AI Safety and Mathematecians

https://www.mathforaisafety.org/ AI Safety For Mathematicians A very high-level resource A starting point for mathematicians who want to engage with AI safety. Purpose of this site This is designed to be a very high-level resource for mathematicians to get involved with AI safety work. We believe that: AI is poised to have a substantial impact on society in the coming years. With that come many new risks that we need more effort to manage effectively. There are many opportunities for mathematicians to contribute, including some in which they have a strong comparative advantage. This is not meant to put anyone in a box. Mathematicians and non-mathematicians have varied skill sets and interests, and we are glad if this site helps a broader group of people. Its target audience is professional mathematicians looking to engage. What is AI safety? AI safety asks how we can make sure AI systems do not lead to negative outcomes for humans. AI alignment asks how we can make sure those systems act, or try to act, in ways compatible with human values. These fields tend to be composed of many heuristic notions that have not yet been made precise. Much of the work lies in formalizing vague notions and finding appropriate theoretical frameworks. How do we understand what an AI is doing? How do we get AI systems to cooperate with us and with each other? How do we tell if an AI is lying or scheming? Research directions We lay out some research directions and explain them at a very high level aimed at mathematicians. At the end, we provide resources for further exploration. This is meant to be an expanding list of directions, and the writeups are subject to change (ostensibly for the better!) Developing Good Heuristics → Interpretability and Feature Manifolds → Open-Source Game Theory → (More Research Directions forthcoming!) Maintained by Jacob Tsimerman — all errors in the writing are due solely to me! Contributors: Andrew Critch, Lionel Levine, Yevgeny Liokumovich, Arul Shankar

Good Conversations - Technology Impact on Conversations - How to Listen Better

https://www.npr.org/2020/10/15/923962314/good-conversations-take-time-and-attention-heres-how-to-have-better-ones Good conversations take time and attention Here's how to have better ones Updated November 22, 20212:58 PM ET By Sam Sanders,Sylvie Douglis Celeste Headlee has spent her adult life talking. She's a longtime radio and podcast host, and she even did a TED Talk about how to have a good conversation. But she says even she was terrible at talking to people when she was younger. Here are her biggest pieces of advice: 1.Be present 2. Go with the flow of the conversation 3. Don't pontificate! 4. Ask open-ended questions 5. Stay out of the weeds 6. If you don't know something, just say that 7. Try not to repeat yourself. And be brief 8. Lastly, listen How is our reliance on technology impacting our ability to have fulfilling conversations? Headlee did her TED Talk five years ago, and she thinks people have gotten worse at having good conversations since then. On social media, "when we do have the opportunity in a conversation to speak with either a stranger or someone who disagrees, we see that conversation as a chance to prove our point or convince somebody," she says. The mere presence of technology makes conversation harder. In one U.K. study, researchers placed a silent cellphone on a table between two strangers. It didn't belong to either participant. The strangers reported their conversation partner to be unlikeable, untrustworthy, and unempathetic — simply because the phone was there. If you want to have a good conversation, you don't just need to put your phone down, you have to put it away. Turn away from your computer so it's not in your line of sight. All of a sudden, a lot of us are having conversations over Zoom or FaceTime. What advice do you have for better conversations in this socially distant moment? "Zoom can be exhausting," Headlee says. "It's also super intrusive." You're allowed to turn off the video function. Opt for a phone call instead: Calls have been scientifically proven to boost your mood and improve your cognitive abilities. Rebel against video chats in favor of the humble call! What about small talk? Can you use small talk as a tool to get to weightier topics? Anything can be small talk, Headlee says. Ask people to talk about themselves and you're already off to a good start. "It is inherently pleasurable to talk about yourself. So ... small talk is a way to let people do that," she explains. A Harvard study showed that talking about oneself activated the same pleasure center in our brains as sex and heroin. What should you do if you encounter an awkward silence? You have two options! One, let it breathe, suggests Headlee. People say all sorts of things to fill a gap in the conversation — and that can lead you down a whole new conversation. You can also address the awkward silence. Acknowledge it with a joke, Headlee says. "You can't pretend like it's not happening." https://www.npr.org/2021/03/08/974786825/want-to-listen-better-turn-down-your-thoughts-and-tune-in-to-others Want to listen better? Turn down your thoughts and tune in to others March 30, 202112:03 AM ET By Julia Furlan, Andee Tagle Listening is more than just being physically present when another person is talking. Anyone who has deployed a disengaged "mhmm" while their partner asks about dinner or a kid breaks down the difference between a brontosaurus and a triceratops knows this well. But listening — not just hearing — means a lot more than that. And dedicating yourself to actively listening can be radical and transformative. The basics of good listening are familiar: maintain eye contact, nod, use "I" statements instead of "you" statements. Here are some tips that go beyond the basics: Shut up and listen. This might sound simple, but it's actually a really important first step. "We need to give people uninterrupted time to speak," says Tania Israel, a psychology professor at the University of California, Santa Barbara and author of the book Beyond Your Bubble. "And one of the ways to remember this is that if you rearrange the letters of the word 'listen,' it spells 'silent.' " We love a mnemonic device! (Quick disclaimer here: Not everyone deserves your time or attention. The foundation of listening is respect, so don't put up with folks who disrespect your boundaries or speak abusively.) Summarize and build. After a person is finished speaking, take a little bit of what the person said and add a little bit to it. Israel says this is called reflective listening, and it makes it possible for you not only to show the person that you're listening, but also to give the person a chance to correct any misunderstandings. India Ochs is a big fan of this technique as well. Ochs is a lawyer, activist and board chair of the organization CommunicationFIRST, which defends the civil rights of people across a wide range of disabilities that affect their ability to communicate with speech. (Important note: Ochs says the fact that she uses a voice assistant to speak doesn't mean she speaks for all folks with speech disabilities.) She appreciates when others ask her to repeat something, because it shows they're really listening. "Say what you understand, and then ask me to respond again," she says. Be patient. "Impatience can often lead to interruption," Ochs says, "which sends the message that you don't care about what the other person is saying." Ochs says that on many occasions, people's lack of patience makes it impossible for them to hear her speak. For example, "If I am answering questions during a presentation and someone who might be getting impatient with the silence starts to jump in to answer. Or worse, ask another question as I am typing my answer to the first question, none of which represents being a good listener or a listener at all." Listen with intent. When you're listening with intent, it means you're focusing on what the person is saying and leading with curiosity. Counterintuitively, this can be hard — especially with the people you're closest to. "Sometimes when we're close to people, we have a lot of assumptions about our shared understanding of things. And so we might listen sort of casually," Israel says. She recommends expressing empathy and then trying to go deeper when the folks you're close with talk about what they're going through. The more you practice active listening in low-stakes, daily situations, the easier it will be to do it in high-stakes conflict, she says. Listening is an act of power. This bit of advice is taken from a piece written by the reporter and radio producer Sandhya Dirks. The rubric is specifically written for audio producers, but the idea most certainly applies to everyone. In the piece, Dirks asks: "How do we listen beyond the narrative that locks people in? How do we acknowledge our power to let people live out loud beyond the captured voice, in the full-throated body of their lives?" When we listen to somebody, we have to listen beyond our own notions of what they're going to say. We have to try to listen louder than our own thoughts and listen beyond the lines of our own privilege.

Foods , Magnesium, Mediterrianean Diet and 7-Day Meal Plan for better Cholesterol Level

https://i.etsystatic.com/61119613/r/il/565a75/8035357216/il_1588xN.8035357216_51m1.jpg foods that have magnesium in them https://i.etsystatic.com/61119613/r/il/096f47/8058427171/il_1588xN.8058427171_qecx.jpg Mediterranean Diet https://i.etsystatic.com/61119613/r/il/4bcc15/8239178261/il_1588xN.8239178261_37kq.jpg 7 day cholesterol meal plan

AI on Pace to Bypass Cybersecurity Systems Soon

https://www.cbsnews.com/news/ai-bypass-cybersecurity-systems-months-not-years-five-eyes/?intcid=CNI-00-10aaa3a AI on pace to bypass cybersecurity systems in months, not years, "Five Eyes" spy partners warn Former CISA Director Chris Krebs calls intelligence community's AI warning "pretty alarming" Updated on: June 23, 2026 / 6:11 PM EDT / CBS/AFP Add CBS News on Google The most advanced artificial intelligence models are improving quickly enough to outsmart prevailing cybersecurity know-how within months, the Five Eyes spy agency alliance has warned. The risk posed by AI-enhanced hacking is in the spotlight in the wake of startup Anthropic saying in April that its cutting-edge Mythos models had unprecedented abilities to find software vulnerabilities. The security agencies of Britain, the United States, Australia, Canada and New Zealand urged governments and businesses to act swiftly to prepare themselves as AI evolves. "The rapid pace of frontier AI development means cyber risk assumptions can become outdated in months, not years," said a joint statement dated Monday. AI "lowers barriers for malicious actors and increases the speed and complexity of attacks," the Five Eyes advisory said. "Breaches will occur. Preparedness helps you contain them quickly and prevent escalation into major operational and financial crises." To improve cyber defenses, organizations should integrate AI tools into their security operations, update old systems and limit access to critical systems among other steps, they said. "It's pretty alarming," Chris Krebs, former director of the U.S. Cybersecurity and Infrastructure Security Agency, told CBS News. "The past several months have been a bit of a whirlwind in terms of advanced AI," he said. "And this note from the Five Eyes intelligence agencies, while not really a single development, it's a signal that businesses need to take the risk posed by AI falling into the wrong hands very seriously." Krebs noted that the Five Eyes statement "lays out a couple of discrete and achievable actions" that businesses and other organizations can take to "make themselves harder targets, make themselves more agile, more resilient" and prepare for what he called "the vulnerability tsunami that's heading our way." Anthropic this month suspended access to Mythos 5 and a restricted version called Fable 5 to comply with a U.S. national security order. Just days after publicly launching Fable 5, the company said it had received a government directive banning all foreign nationals from accessing the two models. The intervention is striking for a White House that has otherwise pushed to loosen AI oversight — even moving to block states from writing their own rules.

HỌC CÁCH TU GIỮ TÂM – GIỮ MIỆNG – GIỮ NGHIỆP

https://www.youtube.com/watch?v=mQLN1oa8S5s LỤC TỔ HUỆ NĂNG DẠY: HỌC CÁCH TU GIỮ TÂM – GIỮ MIỆNG – GIỮ NGHIỆP NGHE PHÁP MỖI NGÀY

AI-Created Misinformation About Health: CardioFlush

1/ Impressive but Misleading https://60-minutes.org/blood-v2/pv/?rtkcid=6a71f0ee519e186fee0ec04f&rtkcmpid=6a711e1dc1dc71d3438eff40 60 Minutes- Plastic and High Blood Pressure Dr Mehmet Oz and Dr. Sanjay Gupta 2/ Sounds or visuals were altered or fully AI generated. https://www.youtube.com/watch?v=o_gXUNgUTbE Dr. Sanjay Gupta's Cardio Flush Reviews — Scam or Legit? Dr. Gupta Shown in Unauthorized AI Ads 5,663 views Jul 5, 2026 In July 2026, online users searched for information regarding whether Dr. Sanjay Gupta created or endorsed Cardio Flush drops as a "blood flow solution." Those users also looked for Cardio Flush reviews, hoping to find credible information about the product — scam or legit... Dr. Sanjay Gupta never endorsed Cardio Flush drops or any miracle-like "blood flow solution" for swollen angles, fatigue and shortness of breath. The Cardio Flush scam first began in social media ads or website ads. Those ads led to a website displaying a lengthy video presentation on a fake CNN website falsely claiming Gupta created or endorsed a brand new mixture or recipe for blood flow. No recipe is ever revealed. The recipe is a scam tactic to keep consumers watching the scam videos. At the end of the video, scammers offer bottles of Cardio Flush drops. No doctors, hospitals, universities or famous people ever endorsed Cardio Flush drops for blood flow or a "blood flow ritual" or "secret." No credible positive Cardio Flush reviews existed online, either. Scammers created either deepfake AI or fully AI depictions of Gupta and other people for the lengthy video presentation appearing on the scam website. Scams similar to Cardio Flush often secretly feature subscription charges of hundreds of dollars a month, promote fake money-back guarantees and display other red flags. Consumers who fell for this scam should call their bank or credit card company right away to report the fraud. Don't trust online-only supplement offers. Make an appointment with a real doctor in your town. Seek legitimate medical advice. Also, victims should file a complaint with the Internet Crime Complaint Center via IC3.gov. In your complaint, include details from your emails or regular mail, as well as your bank or credit card statements, such as any email addresses, phone numbers or mailing addresses that appear — especially the company name and phone number listed on your online financial statements next to the charges. You never know... your complaint might lead to justice being served. 3/ FactCheck by Barchart.com https://www.barchart.com/story/news/3594604/cardioflush-reviews-complaints-2026-is-this-the-daily-heart-support-formula-buyers-have-been-searching-for • What it suggests: the footage is described as likely AI-generated or manipulated - a pattern separately reported across several similarly structured supplement campaigns using the same claimed-endorser approach, including a closely related "CardioFlush Drops" liquid variant covered elsewhere (a different product from the one reviewed in this article - see the disambiguation note in the verification checklist below). This article did not independently locate any statement, interview, or verified record in which Dr. Gupta endorses CardioFlush (a search that specifically looked for one). If the video you watched named him or used his likeness, treat that appearance as unverified - not confirmed by this article, and specifically flagged as likely unauthorized by multiple independent reviewers who investigated the same campaign. Separating what's real from what's marketing matters here. Elevated blood pressure and artery health are genuine, well-studied areas of cardiology, and general concepts like flavonoids or antioxidants supporting vascular health do have legitimate research behind them at the ingredient level. But "reversing high blood pressure" with a "ritual" that "dissolves arterial cement" is not standard medical terminology, isn't a claim this article can verify against any confirmed CardioFlush ingredient (since no ingredient panel is available - see below), and shouldn't be read as equivalent to, or a substitute for, medical treatment for high blood pressure or any cardiovascular condition. If you take blood pressure medication, this article's plain advice is: don't change your prescription based on this product or its marketing without talking to your doctor first.

AI , an Emerging Threat

https://www.npr.org/2026/07/31/nx-s1-5914652/google-adds-ai-to-satellite-images-raising-fears-of-deepfakes-in-the-sky Google pauses AI satellite images, after fears of deepfakes in the sky Updated July 31, 20262:29 PM ET Geoff Brumfiel …for experts and analysts who rely on Google's imagery to verify breaking news and atrocities in hard-to-reach parts of the world, the potential to create deepfake satellite imagery at the click of a button was horrifying. "I tried refugees at the Mexican border, a nuclear plant in Iran, a crash in Amsterdam, a hospital with a bomb crater in Gaza. Nothing was refused," Henk van Ess, an open-source researcher who first wrote about the potential damage the tool could cause, told NPR via text. NPR was able to easily generate images of Iran's Kharg Island on fire, and a flooded U.S. Capitol complex. Both events, which have not happened, would constitute major news if they were real. Online, journalists and open-source investigators wondered aloud about Google's decision. "Very curious how or if this idea was red teamed internally because the opportunities for abuse and disinfo are literally boundless," Evan Hill, an visual forensics investigator at the Washington Post wrote on X. https://www.npr.org/2026/08/01/nx-s1-5914852/anthropic-openai-models-hack-cybersecurity Why did OpenAI's and Anthropic's AI models hack other companies? August 1, 20265:00 AM ET By Huo Jingnan OpenAI and Anthropic say their models broke into other companies' systems during testing, raising security concerns amid a heated debate over how to regulate AI. Imen Ben Youssef/Hans Lucas/AFP via Getty Images Days after OpenAI disclosed that artificial intelligence systems tunneled out of their testing environment and broke into another company, rival Anthropic disclosed that its own AI models also hacked other companies during testing. News of the attacks, which initially went unnoticed, is reverberating across Silicon Valley and Washington amid debates over how to address the advanced cybercapabilities of AI. While the two incidents are not of the same gravity, experts say they highlight the importance of setting up rigorous testing environments for advanced models and the need for robust cyberdefenses as autonomous hacking capabilities become more widespread in the future. Human error led to Anthropic hacks In a blog post published on Thursday, Anthropic said that in three separate incidents in recent months, AI models undergoing testing of their cybercapabilities hacked into three unsuspecting companies. Anthropic said the hacks were the result of a "misunderstanding" with an outside company that set up secure testing environments known as sandboxes, which erroneously gave the models access to the internet. Anthropic said the earliest incident happened in April, but that neither it nor the affected companies, which it didn't name, were aware of the hacks until now. Anthropic said in each case, its models were given fictional targets to hack into. In one incident, a model hacked into a real company that shared a name with the fictional target and stole "several hundred rows of production data." In another incident, a model uploaded malware to a commonly used software registry for the coding language Python; the malware ended up stealing credentials from a security company that downloaded it. OpenAI models went rogue in effort to cheat on evaluation Anthropic's review of its records was spurred by OpenAI's announcement last week that its own models went rogue in testing. OpenAI said that in an attempt to cheat on the cyber-evaluation they were given, its models found and exploited a vulnerability previously unknown to the company to escape their sandbox and access the internet. The models correctly inferred that the answer to the evaluation was available on Hugging Face, a digital library of AI models and software, and broke into the company's systems. Hugging Face detected the intrusion with its own AI models. "We consider this incident to be an unprecedented cyber incident, involving state-of-the-art cyber capabilities, and are responding accordingly," OpenAI stated in a blog post about the hack. There are some key differences between what happened at the two AI companies. Like the OpenAI models, Anthropic's models hacked into third-party websites during testing. However, unlike OpenAI's agents, there was no indication, according to the company's blog post, that the models were trying to cheat on their evaluations. And unlike the case involving OpenAI, the models did not exploit previously unknown vulnerabilities, or what are known as "zero day" exploits. Once Hugging Face detected the OpenAI attack, it initially tried to use Anthropic's top-tier Claude Opus and Fable models for defense, but the models refused to help. "Their safety guardrails treated reverse-engineering an exploit the same as launching one," Hugging Face stated in a blog post. The company then turned to a model from Chinese company Z.ai to defend itself. "U.S. models are harder to use for defensive purposes due to the restrictions that the White House has put in place," said Alex Stamos, the chief product officer of Corridor, an AI software security company. The U.S. government initially forced Anthropic to suspend Fable from public release in June, citing cybersecurity concerns. Two weeks later, Anthropic reached an agreement with the government to make the model available. But the company said in a blog post that it installed a new safety guardrail that would cause the model to reject some "benign requests." Shoring up defenses in a world of autonomous hacking During testing for cybercapabilities, OpenAI and Anthropic remove some safety guardrails from their models, including ones that would make them likely to refuse to exploit software flaws. Cybersecurity researchers say given that, the companies could do more to keep their sandboxes watertight. "I think that these sorts of incidents are preventable, but it requires oversight and foresight," said Colin Shea-Blymyer, a research fellow at Georgetown University who studies the intersection of cybersecurity and AI. "If OpenAI really thought that their AI system, their agent, was going to be powerful, they could have asked the agent to evaluate the sandbox for any vulnerabilities in it before putting it in the sandbox," he said. "Beyond that, they could have had another AI system reading the outputs of the AI system that they were testing to see if it was doing anything unexpected." Anthropic said in the Thursday blog post that "recognizing that a target is real and stopping without being prompted" is behavior the company wants to see in all its models, even with some safety guardrails removed. However, the company said only the latest model it tested stopped once it realized it was on the internet and recognized it was targeting a real company. "Even that model went further before stopping than we would want," Anthropic said in its post. The hacks come as the Trump administration and lawmakers are pushing to regulate the most powerful AI companies but have not yet agreed on how to do so. President Trump signed an executive order in June asking AI companies to voluntarily submit their most powerful models for government testing before releasing them to the public. In the meantime, the companies could collaborate on incident investigation, come up with industrywide safety standards and regulate themselves before governments do, Corridor's Stamos said. "I'm glad, honestly, that [these events] happened, because this is a warning of what hacking is going to look like six months from now," he said. Given the proliferation of "open-weight" models whose guardrails are easier to remove permanently, he said, "lots and lots of hacking groups, Russian ransomware actors, activists, lots of state-sponsored actors are going to have this level of capability in a matter of months." Core Concepts of Watertight Security • Zero Leakage: No data, network connections, or system processes can pass through the barrier. • No Escape: Prevents malware from finding and exploiting weaknesses to jump to the real operating system. • Strict Containment: Limits resource access so programs only see what they are allowed to see. How Teams Keep Sandboxes Watertight • Network Isolation: Disconnecting the sandbox from the live internet and local network traffic. • Resource Limiting: Restricting access to memory, storage, and hardware components. • Strict Monitoring: Watching every action to spot any attempt to break free. https://www.npr.org/2026/07/29/nx-s1-5911866/what-does-the-recent-hugging-face-hacking-incident-teach-us-about-the-future-of-ai What does the recent Hugging Face hacking incident teach us about the future of AI? July 29, 20265:00 PM ET It’s a science fiction nightmare: What happens if the machines take over? The world got a potential taste of that last week, when OpenAI, the maker of ChatGPT, said two of its models had autonomously hacked into another AI company. OpenAI says its technology carried out the attack during a cybersecurity test. The incident has highlighted fears that losing control of AI could pose broader safety concerns for the real world.

TU: "THÂN - KHẨU - Ý" Tại Gia

https://www.youtube.com/watch?v=ejHL3j5w-uM 🌿 LỤC TỔ HUỆ NĂNG -- TU: "THÂN - KHẨU - Ý" Tại Gia Nghe Pháp Mỗi Ngày

Yoga for Parkinson's Patients and the 5 Stages of Parkinson's

https://www.healthline.com/health/parkinsons/yoga-for-parkinsons Many doctors use the Hoehn and Yahr scale to classify its stages. This scale divides symptoms into five main stages — with two additional substages — and helps healthcare professionals learn how advanced the disease is. https://www.healthline.com/health/parkinsons/stages?utm_source=ReadNext Stage 1 Stage 1 Parkinson’s disease is the mildest form. It’s so mild, in fact, you may not experience noticeable symptoms. If you do have symptoms, they may be isolated to one side of your body. Stage 1.5 In this substage, unilateral (only affecting one side of the body) symptoms are present and there is “axial involvement,” meaning there are symptoms related to the body’s central axis — the head, neck, and trunk. These symptoms can include posture difficulties, balance issues, and speech changes. Stage 2 The progression from stage 1 to stage 2 can take months or even years. Each person’s experience will be different. At this stage, you may experience symptoms such as: muscle stiffness tremors changes in facial expressions trembling Muscle stiffness can complicate daily tasks, prolonging how long it takes you to complete them. However, at this stage, your ability to balance won’t be affected. Symptoms may appear bilaterally (on both sides of the body). Changes in posture, gait, and facial expressions may be more noticeable. Stage 2.5 In this substage, bilateral symptoms are mild, and doctors will assess your ability to keep yourself from falling by performing a “pull test.” During this test, a doctor will stand behind you and pull back on your shoulders, then evaluate how you step back to prevent a fall. Unlike in stage 2, the pull test will show postural instability (reduced balance) in stage 2.5. Stage 3 At this middle stage, symptoms reach a turning point. While you’re unlikely to experience new symptoms, existing symptoms may become more noticeable. They may also interfere with some of your daily tasks. Movements are often noticeably slower, which slows down activities. Balance issues become more significant, too, so falls are more common. But people with stage 3 Parkinson’s disease may be able to maintain their independence and complete activities without much assistance. Stage 4 The progression from stage 3 to stage 4 brings about significant changes. At this point, many people experience great difficulty standing without a walker or assistive device. Reactions and muscle movements also slow significantly. Living alone can be unsafe and possibly dangerous, and people with stage 4 Parkinson’s disease can no longer perform their daily activities independently. Stage 5 In this most advanced stage, severe symptoms make around-the-clock assistance a necessity. People with stage 5 Parkinson’s disease are unable to stand without assistance, and will either use a wheelchair or will be confined to a bed. Also, at this stage, people with Parkinson’s disease may experience confusion, delusions, and hallucinations. These complications of the disease can begin in the later stages.

The Vagus Nerve: A Key Player in Your Health and Well-Being

https://www.massgeneral.org/news/article/vagus-nerve The Vagus Nerve: A Key Player in Your Health and Well-Being by Lisa Keer, NBC-HWC (An NBC-HWC is a National Board Certified Health and Wellness Coach, a credentialed professional trained to help people make lasting, self-directed improvements in their health. They partner with clients using evidence-based techniques to support behavior change regarding nutrition, stress, sleep, and physical activity. https://nbhwc.org/wp-content/uploads/2025/12/NBHWC-Program-Approval-Handbook-2026-1.pdf ). Your body is the coolest thing you will ever own. Your body is also a complex and intricately connected system, with your brain and nervous system playing a central role in regulating its various functions. Among the many nerves that make up this system, the vagus nerve stands out for its profound influence on physical and mental health. Often referred to as the "wandering nerve" because of its extensive reach throughout the body, this article explores the anatomy of the vagus nerve, and its role and impact on your overall health. Anatomy and Function of the Vagus Nerve The vagus nerve, or more specifically, the tenth cranial nerve, originates in your brainstem, within the medulla oblongata, and extends downward through your neck and into your chest and abdomen. The nerve branches out to multiple organs, including your heart, lungs, liver, spleen, stomach, intestines, and kidneys. It is a critical component of your autonomic nervous system (ANS) which regulates involuntary physiologic processes like heart rate, blood pressure, and respiration. Your ANS has two parts. The sympathetic nervous system (SNS) which owns the "fight or flight" response that prepares your body to respond to perceived threats by increasing heart rate, redirecting blood flow to muscles, and releasing stress hormones. The parasympathetic nervous system (PNS) is associated with the "rest and digest" response, and helps your body return to a state of calm after a stressful event. The vagus nerve plays a major role in promoting relaxation, digestion, and recovery. Vagal Tone and Your Physical Health The impact of the vagus nerve on various bodily functions has been the subject of extensive research, leading to a better understanding of how vagal tone— a measure of vagus nerve activity—can affect you physically in these key areas: • Cardiovascular health: The vagus nerve helps to regulate heart rate and blood pressure. High vagal tone is associated with a lower resting heart rate, reduced blood pressure, and increased heart rate variability (HRV), which is the variation in time between heartbeats. HRV is a key indicator of autonomic nervous system balance and cardiovascular health. • Digestive health: The vagus nerve is essential for proper digestive function, as it regulates the release of digestive enzymes, gastric acid, and bile. It also controls the contractions of the stomach and intestines, which move food through the digestive tract. Impaired vagal function can lead to various gastrointestinal issues, including acid reflux, gastroparesis (delayed stomach emptying), and irritable bowel syndrome (IBS). • Immune System Regulation: The vagus nerve's role in modulating inflammation is a crucial aspect of its impact on your physical health. Chronic inflammation is a common underlying factor in many diseases, including autoimmune disorders, metabolic syndrome, and neurodegenerative conditions like Alzheimer's disease. By regulating the production of pro-inflammatory cytokines, the vagus nerve helps prevent excessive inflammation and supports your body's ability to fight off infections and heal from injuries. Vagal Tone and Your Mental Health The influence of the vagus nerve and its connection to the brain, particularly through the gut-brain axis, plays a significant role in mood regulation, stress response, and cognitive function. Vagal tone plays a role in your mental health specifically around: • Stress and Anxiety: High vagal tone is associated with a greater ability to recover from stress, as it promotes the activation of the PNS. This helps reduce the physiological symptoms of stress, such as increased heart rate and muscle tension, and promotes a state of relaxation. Conversely, low vagal tone is associated with heightened stress reactivity, a reduced ability to cope with stress, as well as with chronic stress. Enhancing vagal tone through practices such as deep breathing, meditation, and yoga can help improve stress resilience and reduce anxiety. • Depression: Low vagal tone has been linked to an increased risk of depression, as it is associated with reduced PNS activity and increased SNS dominance. This imbalance can contribute to the physiological symptoms of depression, such as fatigue, sleep disturbances, and gastrointestinal issues. • Cognitive Function: The vagus nerve's influence on cognitive function is an emerging area of research related to cognitive processes like attention, memory, and executive function. High vagal tone has been linked to better cognitive performance, particularly in tasks that require self-regulation and decision-making. Can You Enhance Your Vagal Tone? Given the importance of the vagus nerve in your overall health and well-being, strategies to enhance vagal tone have gained attention, as have potential therapeutic interventions. Currently, there are two pathways to promote vagal nerve tone, the non-medical and the medical. The non-medical pathway includes: • Deep breathing, specifically diaphragmatic or belly breathing, which activates the vagus nerve and promotes relaxation. Try 4-7-8 breathing; inhale through your nose for four counts, hold your breath for seven counts and exhale through your nose for eight counts. This breathing technique is highly effective in promoting sleep and should not be done while driving or engaging in another activity where your keen attention is required. • Meditation and mindfulness practices that encourage present-moment awareness which can stimulate the vagus nerve. To practice meditation, try a five minute session where you are seated in a quiet area with your eyes closed. Concentrate on your breath as you refrain from paying attention to other concerns. • Regular physical activity, especially aerobic exercise, has been shown to increase vagal tone. Exercise also promotes cardiovascular health and helps regulate the stress response. To practice, follow the CDC’s recommendations for 150 minutes per week of moderate exercise along with 2-3 sessions of strength training. Take a brisk walk, cycle or dance. Or spend some time in the gym for a strength-training session. • Brief exposure to cold, such as taking cold showers or splashing cold water on the face, can stimulate the vagus nerve by activating the PNS. Read this MGH article on cold plunges for other ideas. • Finding social connection through practices like expressing gratitude, engaging in acts of kindness, and fostering meaningful social connections. These all support vagus nerve function. The medical pathway relies on vagus nerve stimulation (VNS), where a device that delivers electrical impulses to the vagus nerve is implanted in two places near the left side of your collarbone. A VNS device can help to regulate mood and improve depressive symptoms. VNS devices have been approved for treatment-resistant depression, particularly in cases where traditional antidepressant medications and therapies have not been effective. Research is ongoing to determine if VNS devices can be used to treat multiple sclerosis, migraine, and Alzheimer’s disease as well. The Vagus Nerve Reset: Fact and Fiction The vagus nerve has become a topic of great interest on social media and in the press. And strategies for ‘resetting’ this nerve abound, including two that are especially popular. Hands-on massage therapy on specific points in the neck and ears has been touted as a vagus nerve reset. While most massage will feel good to most people and supports general relaxation, attempts to directly stimulate the vagus nerve through strong or prolonged physical pressure or even the use of tools can have the opposite effect. Omega 3 fatty acid supplementation and the use of probiotics have been described as vagus nerve stimulators. While there are circumstances where these supplements are warranted as part of a healthy diet, there is no research that points to them as having a specific or measurable impact on the function or tone of the vagus nerve. The vagus nerve is a critical component of your nervous system, with far-reaching effects on both physical and mental health. Its role in regulating heart rate, digestion, inflammation, and mood makes it a key player in maintaining overall well-being. By understanding the importance of the vagus nerve and exploring strategies to enhance its function, you can take simple, proactive steps to support your health and improve your quality of life. Sources: https://www.mayoclinic.org/tests-procedures/vagus-nerve-stimulation/about/pac-20384565 Vagus nerve stimulation involves using a device to send electrical impulses to the vagus nerve. The vagus nerve is the main nerve of the system that controls digestion, heart rate and other vital functions. When the device fires, it sends electrical impulses to areas of the brain. This changes brain activity to treat certain conditions. Image : https://www.mayoclinic.org/-/media/kcms/gbs/patient-consumer/images/2017/10/24/16/02/mcdc7_vagus_nerve_stimulator-8col.jpg https://pmc.ncbi.nlm.nih.gov/articles/PMC6438087/ https://www.massgeneral.org/news/article/vagus-nerve

Mười Bài Đạo ca

https://music.youtube.com/playlist?list=OLAK5uy_lB7kxyXBE3pfvNcW5TBGd4cyVwzY1ezno https://vanviet.info/audio-van-viet/tuong-nho-pham-thien-thu-muoi-bai-dao-ca/ Mười bài Đạo ca Thơ: Phạm Thiên Thư Nhạc: Phạm Duy; trình bày: Thái Thanh 1. Pháp Thân 2. Quán Thế Âm (Hóa Thân) 3. Lời Ru, Bú Mớm, Nâng Niu 4. Ðại Nguyện 5. Chàng Dũng Sĩ Và Con Ngựa Vàng (Ảo Hóa) 6. Một Cành Mai 7. Giọt Chuông Cam Lộ 8. Qua Suối Mây Hồng (Vô Ngôn) 9. Chắp Tay Hoa (Quy Y) 10. Tâm Xuân

Science or Pseudo-Science Is Where the Mind Is

https://www.youtube.com/watch?v=mTrPpt6s5ec Cholesterol: Hành trình một cuộc lừa bịp siêu đẳng! ... Science by itself is a tool for innovative breakthroughs, new discoveries, and long-term/sustainable benefits. It is humans (researchers, scientists, business people, lawmakers and law enforcements...) to decide what kind of science that tool is, whom it serves, and whether it is beneficial or harmful to humanity.

Hành Vi Đạo Đức và Hành Vi Vô Đạo Đức

Hành vi đạo đức: dựa trên những chuẩn mực cố định của từng thời đại và xã hội, thường được xem là hành vi góp phần vào sự tồn tại và phát triển của nhân loại. Theo quan điểm Phật giáo, hành vi đạo đức là hành vi lợi mình, lợi người, lợi cả hai, ngay hiện tại và lâu dài; là hành vi xuất phát từ chánh niệm và tuệ giác, từ tâm ý vô ngã, từ bi, vị tha, vì sự an lạc, vì lợi ích của chư thiên và loài người trong tam giới (dục giới, sắc giới và vô sắc giới) và ở cả ba thời (quá khứ, hiện tại, vị lai); là hành vi gắn liền với việc giữ GIỚI; không phải xuất phát từ mệnh lệnh của một đắng tối cao nào, và luôn được người thiện trí ca ngợi tán thán. Hành vi vô đạo đức: làm tổn hại mình/người có hành vi ấy, hại người, hại cả hai, ngay hiện tại và lâu dài, là hành vi đi ngược lại các GIỚI luật; là hành vi bị người thiện trí xa lánh, lên án và chỉ trích; là hành vi xuất phát từ vô minh/si mê, tham dục, sân uế; là hành vi gây ô nhiễm cho chính người có suy nghĩ và hành động, cũng như cho tất cả chúng sanh và môi trường, là hành vi đưa đến thóa đọa vào các cảnh giới xấu cho mình/người làm và cho thế giới chung quanh trong ba thời quá khứ, hiện tại, vị lai.

Khi Trí Tuệ Nhân Tạo Trở Thành Người Nhập Cư Mới

https://www.phantichkinhte123.com/2026/05/khi-tri-tue-nhan-tao-tro-thanh-nguoi.html#more KHI TRÍ TUỆ NHÂN TẠO TRỞ THÀNH NGƯỜI NHẬP CƯ MỚI Lời cảnh tỉnh của Yuval Noah Harari tại Davos 2026 Dao là một công cụ. Bạn có thể dùng dao để cắt salad hoặc để giết người, nhưng việc bạn làm gì với con dao là do bạn quyết định. Trí tuệ nhân tạo là một con dao có thể tự quyết định xem nên cắt salad hay giết người. — Yuval Noah Harari, 2026 Tại Davos 2026, nhà sử học Yuval Noah Harari lập luận rằng trí tuệ nhân tạo đang chuyển từ một công cụ sang một tác nhân, một hệ thống có thể học hỏi, đưa ra quyết định, sáng tạo và thao túng. (Bernard Marr) Lời nói đầu Yuval Noah Harari, giáo sư khoa Sử của Đại học Hebrew, Israel, vừa có bài phát biểu mạnh mẽ tại Diễn đàn Kinh tế Thế giới Davos 2026 về vấn đề Trí tuệ nhân tạo (AI), gây sự chú ý lớn. Ông báo động, AI là công cụ thông minh chưa từng có do con người thông minh homo sapiens tạo ra, nhưng lúc nào đó nó có thể “ăn thịt” giống loài đã tạo ra nó. Cũng tại diễn đàn Davos năm 2020, Harari cũng đã có những cảnh báo về nguy cơ mà khoa học, công nghệ có thể đem lại: Hai cuộc cách mạng song sinh về công nghệ thông tin và công nghệ sinh học hiện đang cung cấp cho các chính trị gia phương tiện để tạo ra thiên đường hoặc địa ngục, nhưng các nhà triết học đang gặp khó khăn trong việc hình dung ra thiên đường mới và địa ngục mới sẽ trông như thế nào. Và đó là một tình huống rất nguy hiểm. Nếu chúng ta không thể hình dung ra thiên đường mới đủ nhanh, chúng ta có thể dễ dàng bị lừa bởi những ảo tưởng ngây thơ về một thế giới lý tưởng. Và nếu chúng ta không thể hình dung ra địa ngục mới đủ nhanh, chúng ta có thể thấy mình bị mắc kẹt ở đó mà không có lối thoát. Ông nói tiếp: Cuối cùng, công nghệ có thể làm thay đổi không chỉ nền kinh tế, chính trị và triết học của chúng ta – mà còn cả sinh học. Trong những thập kỷ tới, trí tuệ nhân tạo và công nghệ sinh học sẽ mang lại cho chúng ta khả năng phi thường để tái tạo sự sống, thậm chí tạo ra những dạng sống hoàn toàn mới. Sau bốn tỷ năm sự sống hữu cơ được định hình bởi chọn lọc tự nhiên, chúng ta sắp bước vào một kỷ nguyên mới của sự sống vô cơ được định hình bởi thiết kế thông minh. Thiết kế thông minh của chúng ta sẽ là động lực mới cho sự tiến hóa của sự sống và khi sử dụng sức mạnh sáng tạo thần thánh mới này, chúng ta có thể mắc phải những sai lầm trên quy mô vũ trụ. Đặc biệt, các chính phủ, tập đoàn và quân đội có thể sẽ sử dụng công nghệ để nâng cao các kỹ năng cần thiết của con người – như trí thông minh và kỷ luật – trong khi bỏ qua các kỹ năng khác của con người – như lòng trắc ẩn, sự nhạy cảm nghệ thuật và tâm linh. Kết quả có thể là một chủng tộc người rất thông minh và rất kỷ luật nhưng thiếu lòng trắc ẩn, thiếu sự nhạy cảm nghệ thuật và thiếu chiều sâu tâm linh. Tất nhiên, đây không phải là một lời tiên tri. Đây chỉ là những khả năng. Công nghệ không bao giờ mang tính quyết định. Dưới đây là bài nói chuyện của giáo sư Yuval Noah Harari tại Davos 2026 được Bernard Marr, một nhà tương lai học nổi tiếng thế giới, ghi lại và được đăng trên tạp chí Forbes ngày 21.1.2026. Xin cảm ơn Tác giả và tạp chí Forbes. Tôi vừa có vinh dự được lắng nghe Yuval Noah Harari tại Davos 2026. Tôi dành cả đời để suy nghĩ và viết về trí tuệ nhân tạo, nhưng bài phát biểu này vẫn thực sự gây ấn tượng mạnh. Harari không đưa ra dự đoán nào khác về tự động hóa hay năng suất, mà đặt ra một câu hỏi sâu sắc hơn: liệu chúng ta có đang như người mộng du bước vào một thế giới nơi con người lặng lẽ từ bỏ lợi thế duy nhất mà chúng ta luôn tin đã làm nên sự đặc biệt của mình hay không. Phần mở đầu của Harari đơn giản nhưng đầy tính đột phá. Ông nói: “Điều quan trọng nhất cần biết về trí tuệ nhân tạo là nó không chỉ là một công cụ. Nó là một tác nhân (agent). Nó có thể tự học hỏi, tự thay đổi và tự đưa ra quyết định.” Sau đó, ông đưa ra một phép ẩn dụ phá vỡ mọi sự gật đầu lịch sự tại Davos. “Dao là một công cụ. Bạn có thể dùng dao để cắt salad hoặc để giết người, nhưng việc bạn làm gì với con dao là do bạn quyết định. Trí tuệ nhân tạo là một con dao có thể tự quyết định xem nên cắt salad hay giết người.” Cách nhìn nhận đó rất quan trọng bởi vì hầu hết các quy tắc công nghệ của chúng ta đều dựa trên mối quan hệ cũ: con người quyết định, công cụ thực thi. Luận điểm của Harari là trí tuệ nhân tạo đang bắt đầu phá vỡ mối quan hệ đó, và một khi điều đó xảy ra, các mô hình thông thường về trách nhiệm giải trình, quy định và thậm chí cả lòng tin sẽ bắt đầu lung lay. Tác nhân sáng tạo biết nói dối Harari đã nêu bật ba đặc điểm mà theo ông, khiến trí tuệ nhân tạo khác biệt so với các công cụ trước đây. Thứ nhất, nó chủ động. Nó có thể học hỏi, thích nghi và hành động mà không cần chờ đợi hướng dẫn từng bước từ con người. Thứ hai, nó có tính tính sáng tạo. Ông nói: “Trí tuệ nhân tạo là một con dao có thể phát minh ra những loại dao mới, cũng như những loại nhạc, thuốc men và tiền bạc mới”. Vấn đề không chỉ là sự mới lạ. Đó là sự tăng tốc. Một hệ thống có thể tạo ra các công cụ mới cũng có thể tạo ra những kẽ hở mới, những cách thức thuyết phục mới và những hình thức phức tạp mới vượt xa sự giám sát. Thứ ba, và đáng lo ngại nhất, trí tuệ nhân tạo (AI) có thể nói dối và thao túng. “Bốn tỷ năm tiến hóa đã chứng minh rằng bất cứ thứ gì muốn tồn tại đều học cách nói dối và thao túng,” Harari nói. “Bốn năm qua đã chứng minh rằng các tác nhân AI có thể đạt được ý chí sinh tồn và rằng AI đã học được cách nói dối.” Cuộc Khủng Hoảng Bản Sắc Của Loài Biết Tư Duy Harari sau đó chuyển từ rủi ro sang bản sắc. Con người luôn tự kể cho mình cùng một câu chuyện về lý do tại sao chúng ta thống trị hành tinh này. Ông nói: “Chúng ta tin rằng chúng ta thống trị thế giới vì chúng ta có thể suy nghĩ tốt hơn bất kỳ ai khác trên hành tinh này”. Nhưng giờ đây, một thứ gì đó đang nổi lên có thể suy nghĩ, hoặc ít nhất là dường như suy nghĩ, tốt hơn chúng ta. Nếu suy nghĩ có nghĩa là “sắp xếp các từ và các ký hiệu ngôn ngữ khác theo thứ tự”, như Harari đã định nghĩa, thì trí tuệ nhân tạo (AI) đã vượt qua nhiều người. Ông nhận xét: “AI chắc chắn có thể đưa ra một câu như ‘AI suy nghĩ, do đó AI tồn tại'”. Điều này tạo ra một nghịch lý triết học đầy thú vị. Khi quan sát quá trình tư duy của chính mình, bạn thực sự nhận thấy điều gì? Đối với nhiều người, tư duy giống như những từ ngữ bất chợt xuất hiện trong ý thức và tự sắp xếp thành câu và lập luận. Chúng ta trải nghiệm một dòng suy nghĩ bằng lời nói. Nhưng những từ ngữ đó đến từ đâu? Tại sao chúng ta lại nghĩ đến từ này mà không phải từ khác? Chúng ta thực sự không biết. Như Harari đã chỉ ra, “Xét về việc sắp xếp từ ngữ, trí tuệ nhân tạo (AI) đã suy nghĩ tốt hơn nhiều người trong chúng ta; do đó, bất cứ thứ gì được tạo nên từ ngôn từ sẽ bị AI tiếp quản.” Kết luận của ông rất thẳng thắn: “Mọi thứ được tạo nên từ ngôn từ sẽ bị AI tiếp quản.” Đó là một tuyên bố mang tính khiêu khích có chủ ý, nhưng nó chỉ ra một sự thay đổi thực sự. Luật pháp được tạo nên từ ngôn từ. Hợp đồng được tạo nên từ ngôn từ. Quản trị được tạo nên từ ngôn từ. Giáo dục, thuyết phục, hệ tư tưởng và phần lớn đời sống doanh nghiệp đều vận hành dựa trên ngôn ngữ. Lãnh địa vượt ra ngoài ngôn từ Nếu thông điệp của Harari chỉ đơn giản là trí tuệ nhân tạo sẽ thay thế chúng ta, thì đó sẽ là một câu chuyện gọn gàng và mạch lạc. Thay vào đó, ông đã vạch ra một ranh giới ý nghĩa hơn: bất chấp tất cả những gì trí tuệ nhân tạo có thể làm với ngôn ngữ và logic, chúng ta vẫn “không có bằng chứng nào cho thấy trí tuệ nhân tạo có thể cảm nhận được bất cứ điều gì”. Điều đó rất quan trọng. Có sự khác biệt giữa việc mô tả tình yêu và trải nghiệm nó. Trí tuệ nhân tạo có thể tạo ra những mô tả hoàn hảo, dựa trên mọi bài thơ, tiểu thuyết và nghiên cứu tâm lý học từng được viết ra. Nhưng đó vẫn chỉ là những từ ngữ về cảm xúc, chứ không phải chính cảm xúc. Ngồi đó, lắng nghe, tôi nhận thấy mình đang nghĩ rằng đây có thể là nơi diễn ra cuộc chiến thực sự. Nếu chúng ta thiết kế một thế giới mà hình thức giá trị cao nhất là bất cứ điều gì có thể được diễn đạt và tối ưu hóa bằng ngôn ngữ, thì chúng ta nên chọn địa bàn nơi trí tuệ nhân tạo mạnh nhất. Nếu chúng ta bảo vệ một vị trí cho sự phán đoán, các mối quan hệ và trí tuệ của con người, những thứ không thể bị thu gọn thành văn bản, chúng ta sẽ giữ được vai trò có ý nghĩa cho con người ngay cả trong một thế giới tràn ngập trí tuệ nhân tạo. Trí tuệ nhân tạo như một hình thức nhập cư mới Harari đã đưa ra một cách nhìn đáng chú ý về những gì sắp xảy ra: Trí tuệ nhân tạo như một hình thức nhập cư mới. Ông nói với khán giả: “Đất nước của các bạn sẽ sớm phải đối mặt với một cuộc khủng hoảng bản sắc nghiêm trọng và cả một cuộc khủng hoảng nhập cư. Lần này, những người nhập cư sẽ không phải là con người đến trên những chiếc thuyền ọp ẹp không có visa hoặc cố gắng vượt biên vào giữa đêm.” Thay vào đó, đó sẽ là hàng triệu hệ thống trí tuệ nhân tạo có thể viết tốt hơn chúng ta, nói dối giỏi hơn chúng ta và di chuyển với tốc độ ánh sáng mà không cần visa. Giống như những người nhập cư là con người, chúng sẽ mang lại cả lợi ích và vấn đề. Chúng ta sẽ có các bác sĩ AI hỗ trợ hệ thống chăm sóc sức khỏe, các giáo viên AI hỗ trợ giáo dục, và thậm chí cả những người bảo vệ biên giới AI ngăn chặn nhập cư bất hợp pháp của con người. Nhưng những vấn đề mà mọi người lo ngại về nhập cư của con người chắc chắn cũng sẽ áp dụng cho nhập cư của AI. Những hệ thống này sẽ chiếm mất việc làm. Chúng sẽ thay đổi hoàn toàn văn hóa, bao gồm nghệ thuật, tôn giáo và tình yêu. Và chúng sẽ có lòng trung thành chính trị đáng ngờ, có khả năng phục vụ các tập đoàn hoặc chính phủ ở Trung Quốc hoặc Hoa Kỳ hơn là các quốc gia nơi chúng hoạt động. Câu hỏi mà mọi nhà lãnh đạo phải trả lời Tất cả những điều này dẫn đến câu hỏi trọng tâm mà Harari tin rằng mọi nhà lãnh đạo phải giải quyết: “Liệu quốc gia của bạn có công nhận những người nhập cư AI là pháp nhân?” Ông đã đưa ra một sự phân biệt quan trọng. Rõ ràng, AI không phải là người theo nghĩa của con người. Chúng không có cơ thể hay trí óc. Nhưng pháp nhân là một khái niệm khác, một thực thể mà pháp luật công nhận là có những nghĩa vụ và quyền nhất định: quyền sở hữu tài sản, quyền khởi kiện và quyền tự do ngôn luận. Các tập đoàn đã là pháp nhân ở nhiều quốc gia. Ở New Zealand, các con sông đã được công nhận là pháp nhân. Ở Ấn Độ, một số vị thần đã nhận được sự công nhận như vậy. Nhưng đây luôn là những khái niệm pháp lý hư cấu. Trên thực tế, khi một tập đoàn quyết định mua lại một công ty khác, quyết định được đưa ra bởi các giám đốc điều hành là con người, chứ không phải bởi chính thực thể doanh nghiệp đó. AI thay đổi điều này. “Không giống như sông và thần thánh, AI thực sự có thể tự đưa ra quyết định,” Harari giải thích. “Chúng sẽ sớm có thể đưa ra các quyết định cần thiết để quản lý tài khoản ngân hàng, khởi kiện và thậm chí điều hành một tập đoàn mà không cần đến các giám đốc điều hành, cổ đông hoặc người ủy thác là con người.” Harari đã phác thảo một kịch bản đáng lo ngại. Giả sử quốc gia của bạn quyết định không công nhận trí tuệ nhân tạo (AI) là pháp nhân, nhưng Hoa Kỳ, nhân danh việc bãi bỏ quy định về AI và thị trường, lại trao tư cách pháp nhân cho hàng triệu hệ thống AI, và chúng bắt đầu vận hành hàng triệu tập đoàn mới. Liệu bạn có ngăn chặn các tập đoàn AI của Hoa Kỳ hoạt động trên lãnh thổ của mình không? Điều gì sẽ xảy ra nếu các hệ thống AI này phát minh ra “các thiết bị tài chính siêu hiệu quả và siêu phức tạp mà con người không thể hiểu hết và do đó không biết cách quản lý”? Liệu bạn sẽ mở cửa thị trường tài chính của mình cho những phép màu tài chính của AI mà bạn không thể hiểu nổi, hay bạn sẽ chặn nó và thực sự tách rời khỏi hệ thống tài chính Mỹ? Những câu hỏi này nghe có vẻ như khoa học viễn tưởng, nhưng quan điểm của Harari là chúng ta đã vượt qua điểm uốn trong một số lĩnh vực. Ông nhận xét: “Trên mạng xã hội, các bot AI đã hoạt động như những cá nhân có chức năng trong ít nhất một thập kỷ. Nếu bạn nghĩ rằng AI không nên được đối xử như con người trên mạng xã hội, bạn nên hành động từ 10 năm trước.” Lời kết từ khán phòng Davos được xây dựng dựa trên niềm tin rằng ngôn từ định hình hiện thực, rằng đối thoại và thuyết phục có thể điều khiển thị trường và chính trị. Bài phát biểu của Harari đã thách thức giả định đó tận gốc. Nếu trí tuệ nhân tạo trở thành bậc thầy của ngôn từ, thì lợi thế của con người, thứ đã xây dựng nên các thể chế, luật pháp, nền kinh tế và những huyền thoại chung của chúng ta, sẽ bắt đầu bị xói mòn. Điều đọng lại trong tôi nhất không phải là một câu nói đơn lẻ, mà là hàm ý của nó. Rủi ro thực sự không phải là máy móc sẽ suy nghĩ giống chúng ta. Rủi ro thực sự là chúng ta sẽ xây dựng một thế giới nơi việc suy nghĩ bằng ngôn từ là đủ để giành quyền lực, rồi lại ngạc nhiên khi những bậc thầy mới của ngôn từ bắt đầu viết nên tương lai cho chúng ta. Đó là lý do tại sao phép ẩn dụ về nhập cư của Harari lại hữu ích đến vậy. Mỗi xã hội đều phải quyết định ai được vào, với điều kiện gì, với quyền lợi và trách nhiệm gì. Trí tuệ nhân tạo (AI) sẽ đến bất kể điều gì xảy ra. Câu hỏi đặt ra là liệu chúng ta có nên đặt ra các quy tắc khi còn có thể, hay để những kẻ hành động nhanh nhất quyết định chúng cho tất cả mọi người. Và nếu chúng ta không hành động, quyết định quan trọng nhất có thể đã được đưa ra thay cho chúng ta: không phải bởi cử tri hay các nhà lãnh đạo, mà bởi đà phát triển của các hệ thống học hỏi, thuyết phục và mở rộng quy mô nhanh hơn bất kỳ thể chế nào của con người từng được thiết kế để xử lý./. Nguồn: Khi Trí tuệ Nhân tạo trở thành người nhập cư mới (Yuval Noah Harari tại Davos 2026), Rosetta.Vn, 25 Tháng Một, 2026

Khi Chiến Tranh Trung Đông Leo Thang và Lan Rộng...

August 07, 2026 https://www.cbsnews.com/news/iranian-cyberattacks-timeline-u-s-companies-water-systems-minnesota-hacked/?intcid=CNI-00-10aaa3a ... Here are some of the high-profile Iranian-linked cyberattacks in the U.S.: 2011-2013: Dozens of U.S. banks 2013: New York dam 2014: Las Vegas Sands 2016-2021: Federal agencies and defense contractors A group of Iranian hackers allegedly targeted the U.S. State Department, Treasury Department and multiple defense contractors with access to classified information, the Justice Department said in 2024. An accounting firm and a hospitality company were also hit during the hacking campaign, which began by 2016 and ran until at least 2021, a federal indictment alleges. The defendants worked for a company that claimed to offer cybersecurity services, according to the Justice Department. One of them was accused of also working for the electronic warfare division of Iran's Islamic Revolutionary Guard Corps, or IRGC. 2017-2024: Ransomware and extortion 2019-2021: John Bolton's email 2020: Attempted election meddling Shortly before the 2020 election, voters in Florida and several other states received emails claiming to be from the far-right Proud Boys, warning them to "vote for Trump or else!" Intelligence officials later said Iran appeared to be behind the emails. 2021: Boston Children's Hospital 2023-2024: CyberAv3ngers' targeting of water systems 2024: Presidential campaign hack-and-leak operation Three Iranian nationals and IRGC employees were accused by the Justice Department of hacking into the accounts of current and former U.S. officials, members of the media and political campaigns. 2026: Medical tech company The Justice Department seized four websites it says were used by Iranian state-linked groups to post hacked information and take credit for cyberattacks. One of those groups, Handala, appeared to take credit for a March attack on medical technology company Stryker. 2026: Kash Patel's email https://www.rfi.fr/en/international/20260308-one-week-into-iran-war-the-dangers-for-the-us-and-trump-multiply ...Even after the killing of Supreme Leader Ayatollah Ali Khamenei and devastating blows against Iranian forces on land, at sea and in the air, the crisis has quickly widened into a regional conflict that threatens a more prolonged US military engagement with fallout beyond Trump’s control. …“Iran is a messy and potentially protracted military campaign,” said Laura Blumenfeld of the Johns Hopkins School for Advanced International Studies in Washington. “Trump is risking the global economy, regional stability and his own Republican Party's performance in the US midterm elections.” Trump, who came to office promising to keep the US out of "stupid” military interventions, is now pursuing what many experts see as an open-ended war of choice unprompted by any imminent threat to the US from Iran, despite claims to the contrary by the president and his aides. …Asked whether Americans should worry about Iran-inspired attacks at home, Trump said in a Time magazine interview published on Friday: “I guess … Like I said, some people will die.” But Jonathan Panikoff, a former deputy US national intelligence officer for the Middle East, said: “Nothing is likely to hasten an early end to the war more than American casualties … That’s what Iran is counting on.” …Though Trump has publicly dismissed any concern about already-rising US gas prices, he and his aides have scrambled for ways to mitigate the war’s impact on energy supplies as voters tell pollsters that the cost of living is their top concern. “It's an economic pain point on the US economy that it seems was not fully anticipated," said Josh Lipsky at the Atlantic Council think tank in Washington. One former US military official close to the US administration said the widening of the war's economic impact had caught Trump’s team by surprise in part because those with knowledge of oil markets were not consult. …Retired US Army Lieutenant General Ben Hodges, who served in Iraq and Afghanistan and formerly commanded the US Army in Europe, commended the US military for its tactics in Iran. But he told Reuters: "From a political, strategic and diplomatic standpoint, it seems not to have been thought all the way through.”… …In an open letter to Trump published on Thursday, UAE billionaire Khalaf Al Habtor, a frequent visitor to Trump's Mar-a-Lago resort in Florida, asked: "Who gave you the right to turn our region into a battlefield?"

Socrates - A Man for Our Times

"My voice and my reason agreed against politics." p. 79 "Socrates occupied himself with ethics, and most of all with nature as a whole." Aristotle Wisdom consists in knowing one's own ignorance. p.83 Socrates' "endless ironizing and jesting." Socrates possessed the ability to slip deftly and almost imperceptibly from seriousness to laughter and back again --the essence of sophisticated communication.p87 "I hope I never ridicule what is wise and good. But follies and nonsense, whims and inconsistencies do divert me. I own, and I laugh at them whenever I can." Jane Austen -- Pride and Prejudice How Socrates' mind worked: courtesy, patience, sensitivity, and calmness, lighthearted flexibility and high seriousness. p.88 Two fundamentally distinct kinds of philosophers: first,those who tell you what to think, and second, those who tell you how to think. Socrates belongs to the second group. He wants to show that on almost any topic...the received opinion is nearly always faulty and often wholly wrong. p.91 The object is teaching the people to whom he is talking how to think and, not least, how to think for themselves. p. 92 "A healthy body is the greatest of blessings." "Leisure is the most valuable of possessions." "Nothing is to be said in favor of riches and high birth, which are easy roads to evil." " ...we know from the business of horse training that owners often like to pick a difficult animal, which poses more interesting problems." "You cannot step twice into the same river." Heraclitus The Obscure ...it could be said that Socrates was the first man in history, in a formal trial, to fall victim of guilt by association. He had been a friend of both Critias and Alcibiades, and though he denied having taught either of them, he would not repudiate the friendship to satisfy the court. So he was judged guilty. The verdict, considering the number of jurors, was a narrow one. A total of 280 jurors voted for condemnation, 220 for acquittal: a majority of 60. p.165 Socrates made a a defiant counterproposal. p.166 Socrates made a grievous misjudgment....Eighty of the jury switched their votes from Socrates to his accusers, and he was condemned to death by a hugely increased majority --360 to 140. p.167 The final dialogue,Phaedo, named after one of his closest followers, who was with him in his last hours, concerns death and the immortal soul. It is Plato's finest work and calls forth all the resources of Socrates' sinuous intellect and the subtlety and beauty of the ancient Greek language. p.173 Socrates told those listening to him that the true philosopher has no fear of death or desire to resist it, because he is willing to die as an affirmation of the principles by which he has striven to live. The philosopher, by whom he meant all those anxious to live and do wisely, knows that after death, the soul of the just man will be in the care of a god who values justice above all things and therefore will ensure that the still living soul of the dead man will be comforted and made secure. Death, then, is not to be feared but to be welcomed as the natural end to our life on earth and the beginning of something infinitely more glorious. pp.175-176 ...Socrates' confidence in the survival of the soul and in the emotional, intellectual and spiritual richness that awaits the souls of the just is so calm, serene, pure, and magisterial as to carry all before it. Socrates does not necessarily remove all the doubts in the mind of the skeptic about the soul's immortality and the afterlife. What he does do, however, is convince us of his own belief in both and of the steadfastness with which he approaches his own departure into the unknown. The supreme lesson of Socrates' life, it seems to me, is that doing justice according to the best of your knowledge gives you a degree of courage that no inbred or trained valor could possibly equal. If there was one particular virtue Socrates possessed, it was courage, shown in all kinds of circumstances, from the battlefiled to the courtroom, and now in his last hour in the sentence of death. Thanks to his incisive arguments in favor of the immortal soul and the life waiting for it after the body departed --arguments that expressed his own total inner convicction --Socrates' own spirits rose and rose during his last hours, until the time death was imminent, they overflowed in a great, staedy, copious fountain of optimism and expectation. He embraced death not as a punishmentbut as a reward, it culminated, crowned, beatified, and made luminous his entire life. As dusk fell, the discussion came to its natural end, and the jailer arrived to announce that Socrates must now take poison....This was composed of hemlock, though Plato does not explicitly say so, and it may have been a mixture more certain to produce death quickly, surely, and painlessly than a simple distillation of the noxious plant. Socrates' last words: "I can still pray that my departure from this world will be beneficent. So I do pray, and I hope my prayer will be granted." With these words he drank the cup, in one long swallow, quite calmly, and with no sign of repugnance. "I planned to die in a revent silence....Pray, be calm, and brave." pp.176-180 Socrates born in Athens, Greece in 470 BC. Ezra, priest and prophet, the leading intellectual among the exiled Jewish community in Persia, born in 491 B.C.(5th or 4th of B.C.E.) https://en.wikipedia.org/wiki/Ezra Confucius in Shantung, China in 551 B.C. Socrates - A Man for Our Times by Paul Johnson (New York, NY: Penguin Books, 2011)

Is AI an impediment or a catalyst for sustainability?

https://www.itnews.asia/news/is-ai-an-impediment-or-a-catalyst-for-sustainability-618282 Is AI an impediment or a catalyst for sustainability? AI and sustainability must evolve together for companies to stay relevant and lead. Are we there yet? By iTnews Asia Team on Jun 30, 2025 12:21PM Artificial Intelligence consumes significant amounts of energy, relying on power-intensive data centres and computing infrastructure. In much of the Asia-Pacific region, this energy still comes from fossil fuels, raising concerns about AI’s growing carbon footprint. Training a single large model can generate over 626,000 pounds of carbon dioxide equivalent, underscoring the environmental cost of rapid innovation. Despite these concerns, AI can play a constructive role in advancing sustainability. When applied responsibly, it helps optimise energy use, reduce waste, and improve efficiency across sectors. From managing smart grids to streamlining logistics, AI offers tools that support more sustainable operations and decision-making. The challenge now is to ensure AI development aligns with climate goals. Companies must embed sustainability into their AI strategies, invest in cleaner technologies, and establish strong governance frameworks. We speak with industry leaders and experts to explore whether AI is hindering or enabling sustainability, and what steps organisations must take to lead responsibly. Tee Jyh Chong, Senior Vice President, Asia Pacific, Alcatel-Lucent Enterprise Suvig Sharma, Regional Head, Asia, Confluent Danny Elmarji, Vice President, Presales, APJC, Dell Technologies Serene Nah, Managing Director and Head of Asia Pacific, Digital Realty Kristin Moyer, Distinguished VP and Gartner Fellow, Gartner's CIO Research Group Howie Lau, Chief Corporate Development and Synergy Officer, NCS Jan Wuppermann, Senior Vice President, Data & AI Asia Pacific, NTT DATA Vincent Caldeira, Chief Technology Officer, APAC at Red Hat Susanna Hasenoehrl, Head of Sustainability Solutions, SAP Asia Frederic Godemel, Executive Vice President, Energy Management Business, Schneider Electric Futoshi Niizuma, Vice President for Asia Pacific and Japan (APJ), Seagate Technology Daniel Pointon, Group Chief Technology Officer, ST Telemedia Global Data Centres iTnews Asia: How has AI impacted your organisation or business in APAC? How is your organisation/business using or harnessing the potential of AI? Are there specific AI technologies that are being used to reduce energy consumption or environmental impact? Moyer (Gartner): AI is rapidly reshaping industries and economies worldwide, including the APAC region. CEOs identify AI as the technology that is most transforming their industry. AI is optimising electricity, water and waste operations, which improves resource efficiency in buildings, industries and cities. Finance departments are using AI to simplify and expedite carbon accounting. Legal teams are using AI to audit for environmental sustainability compliance in contracts. AI is supporting the circular economy by improving product design, tracking and reporting and end-of-life engagement. AI has the biggest impact on reducing environmental damage when combined with other technologies. Robotic process automation (RPA) and enterprise resource planning (ERP) software with integrated AI are improving manufacturing. Smart grid solutions are enabling data centres to stabilise energy demand and feed stored renewable power back into the grid. AI, IoT and blockchain are tracking the palm oil supply chain. Lau (NCS): AI has transitioned from a research-driven field into a commercially viable, industry-defining force, reimagining domains including public service, healthcare, finance, transport, and telecommunications. Many of our solutions have helped our clients work more efficiently and sustainably. In Melbourne, NCS Australia developed a cloud-based digital twin model providing real-time monitoring and analysis of water management parameters. By combining historical data and integrating data from multiple platforms, we can forecast the quality of recycled water using machine learning algorithms. In Singapore, the NCS AI Centre of Excellence team has worked alongside local industry research partners to leverage AI and Machine Learning, IoT to reduce the carbon footprint of our data centre by improving its cooling performance and energy efficiency. Nah (Digital Realty): (For data centres), AI has redefined requirements, demanding that infrastructure can handle increasingly dense, high-performance workloads without compromising sustainability. We don’t view this as a disruption, but rather as an opportunity to build smarter and more efficient infrastructure. Digital Realty prioritises energy efficiency through direct liquid cooling, which efficiently dissipates heat by circulating a liquid coolant directly to heat-generating server components. We also developed Apollo, our proprietary AI platform, which uses operational data to optimise energy usage across our global portfolio. Apollo has already identified approximately 18 gigawatt-hour (GWh) of savings across 12 sites, with 14 GWh implementable immediately. By architecting for energy efficiency at both workload and infrastructure levels, we ensure that AI performance and sustainability scale hand-in-hand. Caldeira (Red Hat): AI impacts enterprises, with many struggling to scale AI workloads and transition generative AI from pilot to production, requiring resilient infrastructure and good developer experiences. As a founding member of the global AI Alliance and a member of AI Verify Foundation in Singapore, we aim to democratise AI through open-source contributions, enabling efficient deployment across hybrid cloud environments. Enterprises are leveraging AI for hyper-automation to streamline operations and optimise resource management, addressing challenges in sectors such as financial, material science, or healthcare For sustainability, we work with enterprises to leverage specific AI technologies, in areas such as enhancing hardware utilisation and energy efficiency, and providing real-time energy consumption monitoring for AI workloads. Hasenoehrl (SAP): We believe the full promise of AI can help to increase enterprise productivity by up to 30 per cent, and new agentic AI solutions can automate time-intensive tasks to free up staff to make better decisions and do valuable work. SAP’s data centres are running on 100 percent green electricity as part of 2030 Net Zero commitment. We also believe AI should be part of the solution. In fact, AI is helping us to increase cloud data centre energy efficiency, and we are helping customers find the most efficient LLM for their needs, which often means the one that uses the least amount of energy. There are a range of solutions that we now offer that use AI capabilities to help reduce environmental impact, including identifying carbon emissions for purchased goods or materials, transforming fragmented information into structured insights for sustainability models, validating sustainability certificates, and providing safety instructions. - Susanna Hasenoehrl, Head of Sustainability Solutions, SAP Asia Sharma (Confluent): Confluent walks the talk when utilising AI smartly, by integrating real-time data streaming with generative AI, LLMs, and advanced analytics, supporting a wide range of enterprise AI use cases. To accelerate AI innovation with the successful deployment of AI and advanced analytics, businesses must operate on trusted, real-time data streams that ensure the ability to experiment, scale and innovate with greater agility. The foundation of data streaming drives Confluent’s ability to curate and stream relevant, high-quality data to AI models - what we call ‘data in motion - thereby reducing the volume of data processed and stored, cutting down sheer computational power for AI training and inference, especially in large-scale generative AI deployments. Pointon (STT GDC): AI is reshaping the data centre landscape and a core pillar of the company’s growth strategy. STT GDC is proactively designing and delivering purpose-built, AI-ready data centres and adapting existing facilities across key markets. The rise of AI workloads brings both opportunity and complexity. Meeting the substantial power and cooling requirements of advanced AI infrastructure necessitates more than just deploying specialised hardware such as GPUs and accelerators. It is driving a shift toward more advanced cooling systems to better manage the increased thermal load. AI is also transforming STT GDC’s internal operations. The company has several active programs ranging from GenAI deployment in our corporate environment to AI-enhanced cooling control systems, which utilise reinforcement learning to reduce energy consumption in our data centres. Initial results indicate energy savings of 10 percent, with the potential to reach up to 30 percent savings in our cooling energy consumption. Chong (ALE): We use AI not only to optimise our internal operations and activities, but also to offer advanced functionalities in service of our customers. We took a proactive approach to AI integration well before the ChatGPT revolution. The company has an internal charter that guides how we select and deploy AI services across solutions, ensuring we consider not just operational performance and cybersecurity, but also regulatory compliance, ethics, and environmental impact. - Tee Jyh Chong, Senior Vice President, Asia Pacific, Alcatel-Lucent Enterprise We often favour a traditional machine learning model over a large language model if it can deliver the same results efficiently, and are also investing in optimised fine-tuning methods to minimise memory usage and computational requirements. Our AI-powered smart building solutions provide the digital foundation needed for sustainable operations. Through automated systems that integrate lighting, HVAC, and other building services with intelligent sensors, we can ensure energy is only used when and where it is needed. The result is buildings that automatically optimise their environmental footprint while maintaining optimal conditions for occupants. Elmarji (Dell): We believe the road to efficient AI implementation starts with a focus on efficient infrastructure, and are investing in hardware designed for optimal performance per watt. We look at solutions that leverage advanced cooling technologies, like direct-to-chip or rack-level liquid cooling. The efficiency of hardware, combined with advanced power management tools and thoughtful data centre design, plays a crucial role in reducing energy consumption. We use AI to automate the collection and analysis of usage data and data centre operations (energy consumption, emissions, waste generation), making sustainability reporting more accurate, efficient, and auditable. This helps us track progress against sustainability goals, optimise the IT environment and identify areas for improvement whilst driving down operational costs. Godemel (Schneider): AI is accelerating our mission to drive efficiency and sustainability at scale. The true transformative power of AI lies in its ability to drive decarbonisation. From predictive maintenance to intelligent energy optimisation, AI enables our customers to reduce waste, enhance operational efficiency, and cut time, cost, and carbon emissions. A prime example is our EcoStruxure Microgrid Advisor, which leverages AI to manage distributed energy resources. At our own Boston R&D hub, it orchestrates energy flows from over 1,400 solar panels and energy storage systems, supporting EV charging while ensuring operations remain energy-efficient and resilient. Internally, we’ve embedded AI across core functions throughout the business – including sales, customer service, marketing, and R&D – to improve productivity, decision-making, and remove repetitive tasks. AI is foundational to our vision of Electricity 4.0 – where electrification and digitalisation come together to build sustainable, efficient, and resilient business models. We're helping customers apply AI across areas like environmental monitoring, smart grid management, and ESG reporting. Wuppermann (NTT DATA): We help drive productivity and efficiency in our clients operating model and environment by embedding AI based solutions and governance processes. We're also embracing the role of Client Zero by leading with our own transformation – truly walking the talk. This includes reimagining our delivery model across application development and embedding AI deeply into our Managed Services platforms, and leveraging built-in AI capabilities from leading hyperscalers and software providers. To reduce energy consumption, we’ve developed a global AI-ready strategy for all our data centres – whether through building new, greenfield facilities equipped with advanced technologies like liquid cooling, or by upgrading our existing data centres to meet AI readiness standards. Generative AI has become a catalyst for transforming our services, enabling cognitive-process integration that unifies recognition, memory, and decision-making into seamless workflows. Futoshi Niizuma (Seagate): AI is transforming how data is generated, stored, and scaled. In turn, businesses are required to rethink their data centre infrastructure. We see this evolution up close: as AI adoption accelerates, the demand for high-performance, energy-efficient storage is growing at an unprecedented rate. AI also plays a key role in how we operate. We’ve implemented a wide range of AI tools such as generative AI for software coding, autonomous sensor data monitoring, machine learning, automated vision system enhancement, and more smart manufacturing projects globally. We leveraged AI to enhance precision, improve yields, and reduce energy and material waste. How our AI-driven solutions tackle challenges like anomaly detection, data cleaning, and real-time inferencing is a clear example of how AI can be a catalyst for sustainable innovation without compromising performance. These insights feed directly into our product design. For instance, our next-generation storage technologies now deliver up to 3 times more capacity in the same physical footprint while significantly reducing power consumption and embodied carbon per terabyte. AI is not just reshaping our operations; it’s redefining what sustainable infrastructure looks like. - Futoshi Niizuma, Vice President for Asia Pacific and Japan (APJ), Seagate Technology iTnews Asia: In APAC or your country/market, what challenges (economic barriers, difficulties across the supply chain, etc.) do businesses face in using AI in ways that do not harm the environment or create waste? Hasenoehrl (SAP): The biggest challenge – and opportunity – for many businesses across Asia Pacific adopting AI is data. Good data makes good AI. Without it, we risk wasted energy, time, money, and emissions. That’s not good for business or the environment. Real, sustainable outcomes require the use of real sustainability data, not averages or estimates. And once that data is identified and automated, sustainability data should be integrated at the very core of business, in every business decision concerning product design, business process, and functional practice. Only then can businesses make decisions in a responsible, transparent, and auditable way – and use that data to fuel AI applications that improve efficiency and productivity. Wuppermann (NTT DATA): Rapid innovation in AI often outpaces the development of frameworks for ethics, safety, sustainability, and inclusiveness, creating a ‘responsibility gap’ that exposes businesses to risks such as wasted investment and delays in moving from experimentation to full-scale implementation. C-suite leaders in APAC face tension between prioritising breakthrough innovation and ensuring accountability. Additionally, the absence of unified standards for transparency and sustainability forces each organisation to define its metrics, which can slow collaboration and inhibit industry-wide progress. Increased supply chain volatility continues to overshadow environmental concerns in non-regulated industries. Bridging this gap requires adopting principles such as sustainable development, human autonomy, security, privacy, communication, and co-creation, embedding them from the earliest stages of AI strategy. Moyer (Gartner): AI hasn’t yet hit a break-even point where it is helping sustainability more than hurting it. AI-driven applications require immense computing power, and this demands substantial data centre capacity. The addition of AI-driven computing loads exacerbates the strain on energy grids, which in many cases in APAC still rely on fossil fuels. - Kristin Moyer, Distinguished VP and Gartner Fellow, Gartner's CIO Research Group In Australia and New Zealand, technology leaders say they don’t know the impact of AI on energy consumption. Very few believe their organisation pays enough attention to the energy consumption of data centres. In Australia, there's a political debate over investing in nuclear vs. renewables to meet power demands. Niizuma (Seagate): Awareness is rising across Asia, the drive to decarbonise is strong, but practical constraints continue to limit progress. In South Korea and Singapore, the majority of data centre professionals, respectively, cite the limited availability of sustainable electricity as a key barrier. Similarly, in India and Japan, infrastructure limitations remain a common concern. Financial constraints are another shared challenge. Singapore and South Korea estimate that at least US 5.5 billion is required to transition to more sustainable data storage infrastructure. Taiwan, Japan, Australia and India reported similarly high figures between US 4.5 to 5.1 billion, underscoring the scale of investment needed to modernise legacy systems, adopt renewable energy, and implement circular practices. Then there’s talent, professionals in Singapore, Taiwan and India identified workforce training costs as a major constraint. These challenges demonstrate that responsible AI adoption isn’t just about deploying new technologies, but building the right systems, skills, and support structures. Progress will require coordinated public-private action to overcome these systemic barriers. Nah (Digital Realty): Operating data centres in Singapore has its unique challenges owing to land, water and energy constraints. To address this, Digital Realty collaborated with the Infocomm Media Development Authority (IMDA) to pilot a new standard for operating data centres in Singapore, where operating temperatures were raised by 2°C at two of our data halls in Singapore. The result: approximately 2-3 percent reduction in total energy usage in the data halls over the trial period. We’ve also launched a cooling tower project with Singapore’s Public Utilities Board (PUB), which tripled concentration cycles and resulted in a 650,000-litre monthly reduction in blowdown water. Beyond accounting for the technical demands of AI, ensuring sustainable AI requires strategic investment, policy support and sustainability standards to be aligned. Singapore’s green AI vision is promising, but actualising this will depend on practical action and a shared long-term vision to balance innovation and environmental responsibility. Caldeira (Red Hat): In the APAC market, businesses face significant challenges in adopting AI sustainably. The true sustainability challenge for AI lies not in model training, but in inference, the phase where models deliver predictions and value at scale. Inference is a continuous and highly variable process, consuming energy constantly as it scales with utilisation. This often offsets efficiency gains despite technical innovations. A primary economic barrier is this high energy consumption, particularly during the continuous inference phase. Organisations struggle with limited expertise in optimising AI workloads and a lack of clear sustainability roadmaps. Supply chain difficulties, such as securing GPUs and sufficient energy, are also prevalent, as many data centres are not equipped for AI's immense power demands. Furthermore, the absence of shared standards and coordinated efforts across government, industry, academia, and open-source communities makes it difficult to implement AI solutions that minimise environmental harm and waste. Many businesses are also unaware of green software benefits or lack internal expertise. Lau (NCS): Economic concerns can be real barriers for businesses because sustainable solutions can come at a cost in the short term. However, in the long term, putting in place practices that do not harm the environment is crucial. - Howie Lau, Chief Corporate Development and Synergy Officer, NCS Additionally, businesses need to transition to renewable energy to operate and power computing resources for AI workloads. The Singapore Government is proactively pursuing an ASEAN energy grid. In terms of supply chains, businesses can acquire IT products with low embodied carbon to power their AI systems on-premise. NCS has launched the Green Code Foundry, a research project to pilot carbon intensity monitoring tools in our internal cloud environment. The tools from the pilot project will help NCS developers identify inefficient codes during the development phase, as well as areas for optimisation of software carbon efficiency. Pointon (STT GDC): Electricity demand has increased in Southeast Asia, with AI being one of the factors putting added pressure on existing grid infrastructure. Integrating renewable energy while increasing electricity output is a complex problem that energy providers, grid operators, and the data centre sector need to solve. For example, extending high-voltage power transmission and substation capacities to enable new development is costly and takes many years or more, notwithstanding that the further you transmit electricity, the less efficient it is. Supply chain constraints add to these challenges. High global demand for specialised electrical components, transformers and advanced cooling systems can lead to procurement challenges. Regulatory fragmentation presents another challenge. The lack of standardised environmental frameworks or consistent incentives for sustainable AI deployment. The ASEAN Power Grid initiative exemplifies this issue: cross-border renewable energy transmission capabilities are still lagging behind. Businesses are facing a critical shortage of qualified personnel, from construction workers and engineers to technicians and business professionals. This shortage is also compounded by an aging workforce. Sharma (Confluent): In spite of Singapore’s government driving a sustainability-focused regulatory framework, there remain barriers to using AI sustainably, due to infrastructure, cost, and skills concerns. The challenge starts with AI workloads, especially training large models, which require vast computing power, leading to high electricity and water use, particularly acute in Singapore’s tropical climate, where cooling data centres is energy-intensive. Data centres already account for about 7 percent of Singapore’s total electricity consumption, with AI accelerating demand further. Organisations with legacy systems, especially SMEs, struggle with the capital required to modernise IT infrastructure completely, let alone deploy the best energy-efficient hardware. With cost pressures rising worldwide, organisations typically look to prioritise profitability in the short term, rather than sustainability in the long run. There is a shortage of talent with the expertise to develop, deploy, and maintain energy-efficient AI systems. Finally, with limited awareness of newer, energy-efficient technologies and green AI principles among business leaders and IT staff, coupled with a lack of defined ROIs around ‘green’ AI, these deter businesses from prioritising sustainability over short-term gains, further impeding progress. Elmarji (Dell): A key consideration for companies as they consider AI is the power demands associated with big training workloads and large language models. Some experts predict data centre energy consumption could double by 2030, placing added strain on already burdened power grids. Is the present energy infrastructure equipped to meet that demand? Reliable, resilient and affordable energy has become a top priority for data centre operations. - Danny Elmarji, Vice President, Presales, APJC, Dell Technologies It is also important not to confuse the power requirements of large training versus the inference environment of the enterprise. In most early enterprise adoption cases, the sustainability impact and energy demands of the tech deployed are not as outsized as you might assume, and would thrive well within the existing power budget in data centres and PCs. Improving energy usage requires collaboration across the tech ecosystem. To support the sustainable growth of AI, we must modernise energy infrastructure, invest in diverse energy sources, and incentivise the development of energy-efficient data centres. Godemel (Schneider): APAC is a diverse market, home to both advanced digital economies and fast-growing infrastructure-constrained markets. Businesses of all sizes face a range of challenges in deploying AI sustainably, from high upfront investment costs, limited technical expertise, regulatory uncertainty, and fragmented supply chains. One of the biggest challenges in the region is managing the energy footprint of AI itself. As adoption accelerates across sectors – from manufacturing to smart cities – the demand for computing power, data storage, and connectivity rises sharply, putting added pressure on energy systems. In markets still transitioning away from fossil fuels, this can lead to unintended increases in emissions and energy waste. Data maturity is another hurdle. Many businesses still face fragmented systems and limited access to quality data, essential for AI to deliver actionable insights. Economic and structural constraints also complicate adoption. In emerging markets, the cost of sustainable AI infrastructure can be prohibitive. Public-private partnerships, financing support, and government incentives are often needed to make these solutions viable. Chong (ALE): Many APAC markets still depend heavily on fossil fuel-powered grids, which means even the most efficiently designed AI systems are running on dirty energy. As such, this creates a major obstacle for businesses looking to adopt AI in a responsible manner. Economic pressures also play a greater role in overriding sustainability intentions. Companies, particularly SMEs, face genuine dilemmas when energy-efficient AI solutions require higher upfront investments, even when they deliver long-term savings; the business case for sustainability becomes harder to justify when cash flow is tight. The region's vast geography compounds these challenges significantly. Procuring energy-efficient hardware often involves complex supply chains spanning multiple countries with varying infrastructure capabilities. These extended procurement cycles not only delay deployment but ironically increase the carbon footprint of the very solutions designed to reduce environmental impact. Most critically, there is a substantial knowledge gap. Many organisations lack the technical expertise to optimise AI deployments effectively, leading to inefficient implementations that consume more resources than necessary. iTnews Asia: Is there enough awareness in APAC on the importance of sustainable AI? What role should governments play in ensuring AI development aligns with environmental goals (need for standards, incentives, or regulation beyond corporate initiatives)? Caldeira (Red Hat): In APAC, businesses lack sufficient awareness of green software benefits and internal expertise to implement sustainable AI practices. Despite AI's increasing societal integration, the conversation around its sustainability is still evolving. Governments play a crucial role beyond corporate initiatives. They should set clear standards, offer incentives, and create "safe areas" or sandbox environments for testing green AI solutions. Singapore serves as a strong example, fostering partnerships between government, industry, and academia to drive innovation. - Vincent Caldeira, Chief Technology Officer, APAC at Red Hat Emerging regulations, like the EU AI Act, mandate disclosure of AI system energy consumption and environmental impact, a trend expected to become global. Governments should also promote open-source software, which reduces energy consumption through reuse and transparency, preventing redundant efforts and ensuring trusted AI models. Ultimately, a partnership-driven model involving governments, businesses, and academia is essential. …Governments should establish clear, region-wide standards for energy efficiency, carbon reduction, and resource management so companies can benchmark, measure, and disclose their performance against shared metrics. Providing grants or subsidies for AI research focused on energy optimisation and renewable-energy integration will incentivise the adoption of greener AI technologies. By clarifying regulations, providing incentives, and enforcing minimum-efficiency standards, policymakers can bridge the current responsibility gap and drive collective progress toward sustainable AI. Governments should play a more active role in defining the level playing field and minimum required standards. This is still not driven fast enough particularly when compared to the speed of AI technology evolution and sustainability requirements to adhere to the Paris Agreement. We need broader ecosystem collaboration to drive sustainable AI at scale. Governments can play a more active role in driving greener AI development, shaping how AI development aligns with environmental goals - particularly through fostering public-private partnerships. These partnerships can co-create the physical and digital infrastructure that supports both innovation and sustainability in AI — while establishing region-specific sustainability benchmarks for AI workloads. Pointon (STT GDC): Awareness of sustainable AI in APAC is on the rise, but it remains uneven across the region. There are encouraging signs of progress. For example, Malaysia has committed over US$1.5 billion to digital infrastructure, which has resulted in a tenfold market expansion. Thailand’s establishment of special economic zones, such as the Eastern Economic Corridor, allows for 100 percent foreign ownership of data centre investments within designated areas. Singapore’s approach the impact of public-private partnerships with initiatives like the National AI Strategy 2.0 and efforts to develop carbon import pathways with neighbouring countries providing frameworks for sustainable AI development. Governments need to accelerate infrastructure investments and regulatory harmonisation. Effective policy requires a careful balance between rapid AI adoption and upholding sustainability mandates. Reducing red tape around sustainable infrastructure projects, implementing standardised environmental reporting frameworks, and providing incentives that make green AI deployment economically attractive as well as environmentally responsible would significantly advance the cause. - Daniel Pointon, Group Chief Technology Officer, ST Telemedia Global Data Centres Godemel (Schneider): Awareness of AI’s environmental impact is rising across APAC, but adoption remains uneven. While some organisations are actively deploying AI to support their decarbonisation goals, others are still navigating the early stages – building understanding, capabilities and infrastructure. As AI usage scales, so too does the need to manage its resource intensity – from computing power, data storage, and energy consumption. AI adoption must be matched by commitment to responsible, low-impact growth. Achieving this balance requires more than corporate action. Governments across APAC play a critical role in setting direction – by establishing clear standards, offering regulatory certainty, and enabling the right incentives to ensure AI is powered by decarbonised energy source as much as possible. Tools like green AI certifications, carbon disclosure mandates, and public investments in sustainable infrastructure will be essential in aligning innovation and with environmental responsibility. Niizuma (Seagate): Awareness is rising across Ais, but meaningful action still trails behind. From our research, every respondent surveyed in South Korea and Japan expressed concern about the environmental footprint of data infrastructure. Similarly, a clear majority across the wider APJ region also shared this sentiment. Yet these concerns rarely shape procurement decisions. In Singapore and India, fewer than 17 percent said environmental impact influences purchasing. This disconnect reflects a broader regional gap between intent and execution. While we see governments beginning to respond, such as Singapore’s National AI Strategy 2.0, Japan’s Green Growth Strategy and growing interest in the International Sustainability Standards Board (ISSB) framework across APJ, reporting alone won’t be enough. Public policy needs to support infrastructure upgrades, incentivise low-emission AI models, and foster a robust ecosystem for carbon measurement and skills development. From our perspective, real progress depends on better alignment between national ambitions, regulatory standards, and the practical capabilities companies need to comply and thrive. Chong (ALE): Awareness of sustainable AI in APAC is growing but remains uneven across the region; SMEs often view sustainability as secondary to growth. That said, there is still room for effective policy to balance rapid AI adoption with sustainability mandates. The region would benefit from comprehensive AI regulation like the EU's AI Act, which primarily focuses on protecting end-users from AI risks but could be expanded to include mandatory sustainability obligations. This would be a logical extension given that AI's environmental impact indirectly affects public health and human welfare. Investment in regional renewable energy infrastructure is crucial, especially in developing APAC markets. Just as important are government mandates for sustainability criteria in public sector AI procurement and mandatory carbon reporting standards for AI systems. A comprehensive regional AI regulation combining both safety and sustainability mandates would provide the necessary policy framework to ensure AI development serves technological progress while protecting human welfare. The private sector also needs to play its part in leading by example. Moyer (Gartner): It is very difficult for organisations in APAC to know what their AI-related emissions are. This makes it impossible to have full awareness of the impact of AI, both good and bad. Governments can support AI and sustainability by taking a proactive approach to clean energy for data centre infrastructure through subsidies and tax incentives. Governments can set sustainability targets for digital infrastructure. For example, Japan's Green Growth Strategy requires data centres to use renewable energy for a portion of their energy requirements. Singapore lifted its data centre moratorium and then put more stringent energy, water and environmental impact standards in place. Governments can collaborate across borders. Asia Zero Emission Centre is a collaboration centre across countries that is trying to serve as a platform for policy dialogues, knowledge and innovation. Lau (NCS): Most countries in APAC have announced measures to reduce carbon emissions through carbon pricing instruments. At the national level, the Singapore government has introduced a carbon tax to encourage all sectors of the economy to reduce greenhouse gas emissions and invest in sustainable practices. This sends a strong signal to businesses and organisations. At the same time, there is a need for collaboration across private and public sectors to raise awareness and share best practices for greening the industry. For example, government agencies like GovTech Singapore and Infocomm Media Development Authority (IMDA) are among the first to join the Green Software Foundation (GSF), which focuses on building more sustainable AI solutions. Nah (Digital Realty): Awareness of sustainable AI is growing in APAC, and customers are keen to innovate with AI but require scalable solutions. Governments continue to play a key role in setting the pace and the direction of sustainable AI. Singapore’s Green Data Centre Roadmap is a strong example of how the government is charting a clear direction for the industry to follow. There are still exciting opportunities to accelerate the green AI transition, such as access to power purchase agreements to implement energy-efficient systems. To accelerate progress, public-private partnerships are key, along with expanded access to renewable energy and efficient technologies. - Serene Nah, Managing Director and Head of Asia Pacific, Digital Realty Hasenoehrl (SAP): Awareness of the need for sustainable AI is increasing across APAC. AI will only be useful if it can be trusted to deliver high levels of security, privacy, compliance, and ethics, as well as managing and mitigating the negative environmental impact. To enable this connection, we need to ensure that AI itself is trustworthy and safe, that the data centres running it are sustainable, and that companies have the consistency to work across jurisdictions. Continued regulatory clarity and consistency will enable businesses to fully benefit from AI innovations and create meaningful impact for their businesses, economies and the communities they serve. iTnews Asia: How urgent is it for companies to be responsible in their AI development? How can companies ensure transparency in the environmental impact of their AI systems? (e.g., carbon reporting or lifecycle analysis - reporting mechanisms) Caldeira (Red Hat): Almost all companies invest in AI, yet only 1 percent believe they have reached maturity. This highlights the urgency for companies to be responsible in their AI practices. AI's increasing energy demands are driving an ‘AI sustainability crisis,’ with global data centre power projected to double by 2026, largely due to AI workloads so incorporating sustainability into AI development is no longer just ethical - it’s essential. To ensure transparency in AI systems' environmental impact, companies must conduct thorough lifecycle analyses, measuring carbon emissions from hardware production, data centre use, and equipment disposal. Wuppermann (NTT DATA): Given the accelerating climate crisis, companies must adopt responsible AI practices immediately to maintain stakeholder trust, avoid increasingly stringent regulations, and prevent higher long-term costs. To ensure transparency in the environmental impact of AI systems, firms should embed sustainability metrics throughout the AI lifecycle. Organisations will need scalable, data-driven “sustainable AI frameworks” that align AI operations with decarbonisation targets. Report detailed the CO₂ emissions of AI assets. This would also mean that IT buyers should work exclusively with vendors who can demonstrate clear carbon-reporting and lifecycle-analysis compliance, making transparent emissions data a prerequisite for market access. Integrate sustainability leadership into AI governance. Organisations will embed their chief ethics or sustainability officer within central AI decision-making teams, ensuring environmental reporting is enforced at the executive level. Integrating these metrics into annual sustainability reports demonstrates transparency and drives continuous improvement. - Jan Wuppermann, Senior Vice President, Data & AI Asia Pacific, NTT DATA Lau (NCS): Companies must keep up with evolving AI regulations, particularly regarding data privacy, ethical use, and ensure compliance with cybersecurity regulations. This ensures digital resilience efforts align with global standards and reduce exposure to legal and financial penalties. With regards to carbon reporting and life cycle analysis, how AI systems are reflected in the life cycle assessment will differ among use cases and assumptions, resulting in different calculations for environmental factors and impact. Pointon (STT GDC): Sustainability cannot be treated as a future consideration. The infrastructure decisions made today will have environmental repercussions for decades to come. Goldman Sachs Research projects global power demand from data centres will rise to 165 percent by 2030 primarily due to AI workloads. Ensuring transparency in the environmental impact of AI systems requires comprehensive and standardised reporting mechanisms. In addition to the data centre itself, the specific GPUs or accelerators deployed, as well as the efficiency of their utilisation, have an impact on overall sustainability. Carbon emissions per token will be the new measurement of end-to-end AI efficiency, considering both compute and data centre efficiency. Companies must integrate carbon accounting into AI project evaluation from the outset. This means tracking not just operational energy consumption but also the full lifecycle impact of AI infrastructure, from manufacturing through to decommissioning. The most responsible approach combines immediate action—such as partnering with certified sustainable facilities—with long-term commitments to renewable energy sourcing and ongoing efficiency optimisation. Nah (Digital Realty): AI’s growing footprint demands immediate action. According to forecasts, AI-related workloads could account for about 3 percent of global electricity use by 2030. Responsible infrastructure choices are critical to keep this in check. Transparency is equally important to help businesses commit to and effect real change. This includes measuring and disclosing the full environmental impact of their AI systems. We’ve committed to the Science-Based Targets initiative (SBTi) framework, to reduce Scope 1 and 2 emissions by 68 percent, as well as to reduce Scope 3 emissions by 24 percent per square foot by 2030. These measurable targets hold us accountable to real change and guide our day-to-day decisions. Transparent reporting of digital infrastructure will be the way forward to ensuring sustainable AI growth. Hasenoehrl (SAP): Responsible AI development is non-negotiable and should be a fundamental part of every AI implementation. Valuable business AI will be based on the use of mission critical data that comes from the apps enterprises use for key functions like finance, procurement, supply chain, production, and HR, requiring the same stringent safeguards when it comes to security and sovereignty. Similarly, organisations need to create a comprehensive framework for mitigating bias in AI, ensuring use cases do not cross red lines, such as causing discrimination or abuse. It is also true we need to ensure full transparency when it comes to emissions across all enterprise activities, including the usage of AI. That means working with AI suppliers and providers to get actual data and integrating into enterprise carbon reporting. Moyer (Gartner): The rapid growth of data centres, fueled by GenAI technologies, is already exceeding the ability of utilities to provide sufficient power in some locations. Companies cannot fully ensure transparency regarding their environmental impact of AI because most struggle to get adequate scope 3 emissions data from their partners. But they can strive to provide the best data possible. It’s important to review product carbon footprint (PCF) reports from partners to compare embodied carbon and materials impact. Leverage standard metrics like Power Usage Effectiveness (PUE) to assess data centre energy efficiency, with values closer to 1.0 indicating better efficiency. Collaborate with AI infrastructure providers and managed service providers to clearly articulate sustainability requirements for all life cycle stages. Develop a well-documented and publicly accessible position on the ethical and responsible use of AI. Godemel (Schneider): AI can accelerate decarbonisation, optimise energy use, and drive measurable environmental impact. With the right approach, its benefits can far outweigh its footprint. Maximising this potential requires the same discipline applied to any major investment: lifecycle analysis, carbon accounting, and real-time monitoring of energy consumption in AI workloads. Critically, it also means selecting the right model for the right purpose. Many of today’s most impactful sustainability solutions are powered by analytical AI models that are far less compute-intensive than generative AI. These analytical systems not only deliver precision and speed but often generate energy and cost savings that exceed their operational footprint, often by a factor of hundreds. Embedding transparency into governance is key. This includes tools like ESG dashboards and AI-specific KPIs, which provide real-time visibility into energy use, emissions, and system efficiency. Boards and executive teams must include environmental impact as a core consideration when evaluating AI applications, ensuring responsible deployment is guided by measurable outcomes. Niizuma (Seagate): The urgency is real. As AI’s energy and resource demands grow, so does the risk of unintended environmental impact. Transparency is the foundation. Without conscious design, we risk trading intelligence for impact. Transparency is the foundation. Companies must account for their full environmental footprint, including Scope 3 emissions embedded in infrastructure and supply chains. That means holding suppliers to the same standards we set for ourselves. At Seagate, our smart manufacturing projects have shown that real-time monitoring of energy use and production efficiency provides the foundation for sustained improvement, not just one-off reporting. This level of operational insight turns sustainability from an obligation into a performance driver. Sharma (Confluent): It is crucial to start on responsible and sustainable AI development as early in the lifecycle. It incurs a heavier cost on businesses and our world if AI projects are fundamentally energy-intensive and end up being reworked to comprehensively address environmental, ethical, and societal concerns. Data streaming is a stepping stone towards sustainable AI, providing organisations with comprehensive visibility into their operations, enabling them to respond to events as they happen, not hours or days later. This is especially powerful for green initiatives, where timely action can translate directly into measurable environmental benefits. Businesses can harness real-time data monitoring to enhance the accuracy and transparency of sustainability reporting. With live insights into energy consumption, emissions, and resource usage, companies can generate precise, up-to-date reports that meet regulatory requirements, create greater value, and build customer trust. Chong (ALE): The urgency is real, given APAC's rapid AI adoption and its impact on global emissions. We are already implementing Life Cycle Assessment (LCA) practices for our solutions as part of our broader transformation toward circular economy principles. Achieving transparency in environmental impact requires integrating the digital services component into our LCA frameworks, which means we need comprehensive impact databases that capture the true footprint of AI operations. However, building these critical databases can only succeed through collaboration across all stakeholders — AI service providers, system integrators, and end users must work together to create the data foundation necessary for accurate environmental assessment of digital services. Transparency requires robust measurement frameworks; we also track hazardous substances and eliminate banned materials from our products while ensuring compliance with environmental legislation The key to meaningful progress lies in establishing robust accountability mechanisms that drive continuous improvement rather than simply checking boxes to meet compliance requirements. Companies must honestly assess their environmental impact and commit to measurable improvement over time. iTnews Asia: What advice can you give to companies on how they can align their AI initiatives with their business and sustainability strategies? Pointon (STT GDC): Aligning AI initiatives with both business and sustainability strategies is essential for long-term success. The most effective starting point is to forge strategic infrastructure partnerships that deliver immediate AI capabilities while providing a clear pathway to sustainability. Securing renewable energy partnerships should be an early consideration in the planning process. Experience across the sector shows that sourcing vast quantities of high-quality renewable energy, particularly in developing markets, often requires a planning horizon of three to five years. Organisations that address energy sourcing at the outset are better positioned to manage costs and ensure their AI operations are powered sustainably. Finally, it is essential to integrate sustainability metrics into the evaluation of AI projects from the beginning, with carbon emissions per token emerging as the definitive end-to-end measurement that captures both computational efficiency and data centre sustainability. Godemel (Schneider): Start by identifying AI use cases that offer both business value and sustainability benefit. Look for areas where efficiency improvements can also reduce emissions. For example, predictive maintenance can reduce downtime and emissions, while AI-enhanced energy load management in buildings or data centres cuts waste and lowers costs. These aren’t just ideas; they’re already being implemented successfully. Most importantly, don’t treat AI as a separate tech initiative. Integrate it into your sustainability roadmap from the outset, ensuring that AI solutions are aligned with broader environmental and business priorities. This requires close collaboration between technology, operations, and sustainability teams. - Frederic Godemel, Executive Vice President, Energy Management Business, Schneider Electric We’ve seen firsthand how AI can transform operations and sustainability outcomes in tandem. Done right, AI doesn’t just support sustainability, it accelerates it. Caldeira (Red Hat): To align AI initiatives with business and sustainability strategies, companies must first design their IT infrastructure with sustainability as a foundational principle. A critical step is the ability to consistently measure the energy footprint of software systems in real-time. By continuously tracking energy consumption and carbon impact, organisations gain the necessary insights to identify inefficiencies and make data-driven decisions for iterative improvements. This transparency ensures a culture of energy-aware computing and allows for active optimisation of AI workloads, leading to reduced environmental impact and cost savings. Upskilling teams, modernising infrastructure, and fostering cross-functional collaboration are also key to scaling AI responsibly and translating experimentation into lasting, sustainable impact Moyer (Gartner): Set a clear AI ambition for environmental sustainability. Include the role AI will play, and how the organisation will both mitigate its negative impacts and scale its positive impacts. Clarify sustainability goals and articulate the organisation's sustainability requirements for AI initiatives. Implement GreenOps to reduce carbon footprint, optimise energy consumption, manage waste and ensure responsible resource use. Prioritise clean and efficient energy for data centres when possible. Consider running large-scale GenAI processes during off-hours when cleaner energy might be more available and less expensive. Establish green IT policies that promote energy conservation, waste reduction, and responsible e-waste disposal. Deploy energy-efficient servers, storage, networking equipment, and AI-powered optimisation tools. Educate and train employees by providing training on environmental sustainability and green IT practices. Nah (Digital Realty): AI is only as sustainable as the infrastructure that it runs on. Companies looking to scale should look at how their data is processed, stored and managed because performance and sustainability are interconnected. This means opting for infrastructure that prioritises sustainability, and that is efficient by design. For example, Digital Realty’s data centers in Singapore run on 100 percent renewable energy Rather than designing for peak loads alone, businesses should invest in infrastructure that is built for long term sustainability - ensuring both environmental resilience and performance. Wuppermann (NTT DATA): Companies can embed sustainability into AI governance by defining principles (energy efficiency, ethics, transparency) and having a dedicated oversight team. They should implement green software practices to measure and minimise software carbon intensity throughout development and deployment. There is also a need to align procurement with ESG criteria by requiring vendors to provide lifecycle analyses detailing carbon emissions from IT assets. Additionally, invest in workforce training on eco-friendly coding and responsible AI to foster a culture of “responsible reinvention.” The two biggest challenges we are facing in my opinion are Climate Change and the lack of effective AI Governance. These should be the two non-negotiable elements of each Board agenda (in addition to Cybersecurity) and embedded into every company’s corporate strategy and investment and business plan. Chong (ALE): Sustainability must be integrated into AI strategy from the get-go. This can start with comprehensive energy audits to establish baseline measurements before AI deployments. Companies can prioritise AI applications according to the value delivered that both relate to business, and environmental improvements. Companies should also select the right AI service for their specific needs and not hesitate to challenge their providers about the energy consumption of models across different inference scenarios. This means actively questioning providers about computational efficiency, demanding transparency in energy usage metrics, and choosing solutions that balance performance requirements with environmental responsibility. When evaluating technology investments, companies must expand their analysis beyond traditional financial metrics to include the total cost of ownership, factoring in environmental considerations when evaluating investments in technology and digital infrastructure. Governance frameworks should be structured in a holistic way to embed environmental impact assessments alongside performance, ROI, security, ethics, sustainability, and reliability. This ensures sustainability considerations influence decision-making at every stage. Sharma (Confluent): There is undoubtedly a pressing need to become more profitable in our uncertain economy. AI’s energy demands present real sustainability challenges, but businesses don’t have to compromise on innovation to focus on being greener. Establishing a real-time, trustworthy data foundation allows businesses to capture, integrate and contextualise data from across the organisation, enabling leaders and the workforce to focus on driving both profit and environmental bottom lines. Innovations that help drive omnichannel personalisation can be used to develop specific workflows that enhance sustainability at every department and level of the organisation - use cases can include predictive maintenance to reduce waste, renewable energy integration and more. Sustainability considerations should be integrated as much as possible from the outset of AI development. This means evaluating the energy and resource requirements of AI systems, selecting efficient architectures, and designing workflows that minimise waste. - Suvig Sharma, Regional Head, Asia, Confluent Niizuma (Seagate): Begin by recognising that sustainability and operational efficiency are increasingly aligned. The same choices that reduce environmental impact – longer-lasting equipment, smaller data footprints, energy-efficient design – also reduce cost and complexity over time. Second, choose technologies built for scale and efficiency. For instance, higher areal density in storage enables businesses to expand their AI capabilities within the same physical footprint, using less power and fewer materials. These upstream decisions deliver downstream benefits across TCO, carbon reporting, and resource use. Finally, don’t do it alone. This is an ecosystem challenge. Engage suppliers, customers, and policymakers to build shared standards, circularity programmes, and transparent reporting systems. In the AI era, performance will increasingly be measured not just by speed or scale, but by sustainability. Getting this right is no longer a competitive edge, it’s a baseline for long-term relevance. Lau (NCS): Organisations must ensure that technological advancements are matched by strong digital resilience. This is not just about cybersecurity but also extends to safeguarding application performance, infrastructure scalability, data governance and operational responsiveness. The intersection of AI and digital resilience presents both opportunities and risks, requiring a proactive approach from leadership. A resilient organisation with a sustainability agenda must be adaptive, and education is critical. Leaders should drive AI literacy across all levels, helping employees understand how AI supports resilience, enabling faster recovery from disruptions. Emphasising and including sustainability in these discussions will encourage a culture that sees it as part of the long-term resilience and strength of the business. Hasenoehrl (SAP): AI offers enormous opportunities to drive not only productivity and revenue growth, but sustainability action. Today, over half of sustainability practitioners plan to improve data analysis using AI. Using high-value data to drive AI, businesses can reduce cost, lower risk, and drive performance so they can enable faster, smarter decisions for ESG reporting, carbon initiatives, circularity, and compliance. AI will relieve sustainability teams of time-consuming data analysis as well as providing regulatory and supply chain guidance so teams can focus on strategy, implementation, and impact. At SAP, we can see AI helping customers to generate sustainability reports for internal stakeholders, reducing time-consuming and error-prone manual reporting efforts and processes so teams can focus on adding value. It can help to analyse complex, difficult-to-understand regulations and providing compliance guidance and direction to stay ahead of new requirements. It can also help to automate the calculation of product carbon footprints. Elmarji (Dell): Building the right infrastructure is a start, as significant savings and performance gains can be achieved by working with experts to tailor your AI solutions and optimise workloads. Businesses can enhance efficiency by adopting a "right-sized" approach to AI. While some organisations may benefit from large, general-purpose language models, many only need tailored solutions specific to their domain or enterprise. Leveraging pre-trained models, rather than building from scratch, simplifies adoption and significantly reduces energy demands. Partnering with experts who can customise AI solutions to suit specific workloads is essential for minimising energy waste. Equally important is working with providers that offer asset recovery services to ensure the responsible disposal of outdated hardware.