# Gates: AI regulation harder than nuclear talks > Bill Gates warns that getting countries to agree on AI rules will be tougher than Cold War nuclear negotiations, and calls for immediate safeguards against bad actors. - Title: Gates: AI regulation harder than nuclear talks · NBC News - Summary: Bill Gates warns that getting countries to agree on AI rules will be tougher than Cold War nuclear negotiations, and calls for immediate safeguards against… - Keywords: bill gates, artificial intelligence, global regulation, ai safety, donald trump, china, ai monitoring, cybersecurity, anthropic, openai, ro khanna, gates foundation - Source: NBC News — https://nbcnews.com/news/us-news/bill-gates-global-artificial-intelligence-regulations-nuclear-rcna599646 - Author: Alexandra Marquez - Published: 2026-09-27T14:10:16.224+00:00 - Read time: 3 min - Topics: bill gates, artificial intelligence, global regulation, ai safety, donald trump, china, ai monitoring, cybersecurity, anthropic, openai, ro khanna, gates foundation ## Gates: AI regulation is harder than nukes Bill Gates told NBC News that getting countries to agree on a global AI framework will be "more difficult" than Cold War negotiations to limit nuclear weapons. > This is more difficult than that was. ## Trump rejects AI guardrails President Trump said the US should focus on winning the AI race against China rather than imposing federal regulations on companies building AI models. > WHOEVER WINS AI, WINS! We are leading China, and all others, and will continue to do so. ## Regulation won't slow US progress Gates argued that monitoring systems and safeguards would not handicap the US in competition with China, and predicted Trump would come to agree after hearing from him and others. ## US should lead on global AI rules Gates pointed to the US leading on aviation safety, ozone protection, and nuclear weapons agreements, saying America should do the same for AI as a shared human challenge. ## AI whistleblower warns of extinction risk Jacob Coxon, who recently left Anthropic, publicly warned that people building AI believe it could kill everyone by the end of the decade. > The people building AI earnestly believe that it could kill us all by the end of the decade. ## Top AI CEOs call for slower development Dario Amodei of Anthropic and Sam Altman of OpenAI agreed that AI development should slow to prevent superintelligence that outpaces human thinking. ## Amodei: slow down but stay ahead Amodei wrote that frontier AI companies must slow development only to a point where they still lead China, and urged global coordination between democracies and authoritarian governments. ## Khanna proposes US-China AI treaty Rep. Ro Khanna said the US and China should form a working group to negotiate an enforceable treaty banning self-improving AI and monitoring frontier labs. > I do believe the United States can lead on a treaty that is both enforceable and verifiable, where we ban recursive, self-improving AI, and we have some monitoring of these frontier labs. ## Trump and Xi discuss AI risks Following a summit with Chinese President Xi Jinping, the White House said both countries established a dialogue to exchange views on AI risks and benefits, with another meeting expected by November. ## Bad actors pose immediate threat Gates said the most dangerous near-term risk is people with bad intent using AI to commit fraud, shut down power grids, or scramble bank accounts, empowering small groups to do what only large nations could before. > The most dangerous thing we're facing right now is people with bad intent using AI. ## Monitoring is essential safeguard Gates called for internal monitoring systems in AI labs so law enforcement can see what AI is doing during fraud or cyberattacks, and said these safeguards can be added without slowing development. ## Kill switch alone isn't enough Gates said a kill switch to shut down AI services during emergencies is insufficient without monitoring and records to understand what went wrong and prevent future tragedies. > It's not enough to have a kill switch. ## Gates Foundation expands AI languages The Gates Foundation is working with 60 AI companies to improve AI accuracy in African and other non-English languages, aiming for all languages spoken by a million or more people to match English quality within five years. ## Foundation using AI to guide giving Gates said the Gates Foundation, which will spend over 200 billion dollars over the next 19 years, is using AI to improve decisions on malaria, malnutrition, and other global health priorities. ## Key takeaway Global AI regulation requires unprecedented international cooperation, but immediate monitoring and safeguards can prevent misuse without slowing development.