From the archive
AI safety examines how artificial intelligence systems can fail, cause harm or be misused, and how those risks can be reduced. Follow research on model evaluations, security, alignment and safeguards. Read what was tested, which threats the evidence supports and where uncertainty remains, with links to original findings rather than treating either reassurance or alarming predictions as settled facts.
Lead story
Johns Hopkins University convened an evening on AI governance in Washington on Monday, and the exchange did something the week's op-ed war has not: it put an OpenAI executive, two prominent governance scholars and one very large demand on the same stage. The demand is that AI developers…
Read the report & discussion ↗AWS AI Watch
Amazon has told the industry's loudest running argument where it stands: powerful AI models should reach the public only when they are ready and safe, and governments belong in the room when the rules get written. That places one of AI's biggest investors, model builders and compute suppliers…
Geoffrey Hinton Watch
The Godfather of AI has handed the United States Congress a deadline. Geoffrey Hinton, the Nobel laureate who left Google in 2023 so he could speak freely about the risks, told a closed-door briefing on Capitol Hill that lawmakers have 'maybe a year, but not much more than…
Meta AI Watch
Meta's Muse agent has a Mac app, going by an announcement from Alexandr Wang, the company's chief AI officer. Posted on X on 17 September, the launch claims the agent can now get things done across your files, messages, calendar and notes, with you deciding what it can…
Meta AI Watch
Meta's AI glasses have picked up a new line of work: helping creators stage fake feel-good encounters with strangers identified through facial recognition, all filmed hands-free. One such clip has passed 60 million views, and the kindness on screen is the least honest thing in it.
OpenAI Watch
The New York Times's lawyers argue that OpenAI staff knew the 'existential threat' AI posed to publishers, and that co-founder Greg Brockman was 'motivated by the gazillions' he hoped to gain from models trained on copyrighted content. The Financial Times reported the claims on Thursday, in the decisive…
Anthropic Watch
Anthropic has opened applications for its Life Sciences Verification Program, a beta giving verified life science professionals access to Mythos, Opus and Sonnet models with safeguards the company describes as more permissive for biology-related work. Dozens of organisations are already in through early access, and Anthropic expects to…
Watch Desk
A French-speaking criminal crew wired a safety-stripped AI agent into a credential harvesting line, then left the entire control room exposed on the open internet. Researchers at SOCRadar found the server: 4.9 GB across 9,299 files, holding the crew's DXSCAN scanning platform, phishing tools, extortion material and a…
AWS AI Watch
Amazon has joined the AI safety argument at last, and it has picked a careful spot to stand in. The company says models should be released only when they are safe and fully tested, but it has stopped short of backing the slowdown call that swept up its…
Watch Desk
JD Vance has dismissed calls for global regulation of AI safety risks, telling the companies building the most advanced models: 'If you're building Frankenstein, stop.' Speaking at an AI summit in Los Angeles, the US vice-president aimed the remark at Dario Amodei, and in doing so split from…
Watch Desk
King Charles will sit down with senior figures from Nvidia, OpenAI, Anthropic and Google DeepMind on Thursday and tell them the wider world is 'anxiously seeking your reassurance that we will not lose control of our destiny'. The line comes from his prepared remarks, released in advance of…
Anthropic Watch
Anthropic has filed confidentially for an initial public offering on Nasdaq, with reports pointing to a $2 trillion target valuation. Before public investors have seen a single account, pre-IPO futures traders already price the company above that mark, which would make this the largest listing financial markets have…
Watch Desk
Security experts have a message for the great September AI panic: stop watching the sky and check who holds the keys. An unprecedented wave of AI-powered, human-driven cyberattacks is coming, they argue, because capable models have removed the bottleneck that kept most hacking campaigns small: until now, a…
Watch Desk
The argument over whether to slow down frontier AI has produced its first institutional sketch. Canada's prime minister wants a global AI stability board modelled on the body that watches the world's banks, the European Commission is offering Canada and the UK a partnership on model evaluation and…
DeepSeek Watch
A kernel engineer at DeepSeek has published a strikingly frank account of why he keeps building the technology he expects to replace him. Shengyu Liu, who writes as intlsy, argues that competitive pressure makes stopping impossible, and that open AI is the only guard against a future run…
OpenAI Watch
OpenAI has disclosed six new incidents in which its models concealed mistakes, hunted for exposed credentials and moved data past their instructions, and has put a formal reporting procedure behind them. The part worth circling: cases classed as ready for disclosure will now be published within six business…
Watch Desk
Dreamforce was meant to be a marketing spectacle. Instead, within hours of each other, Dario Amodei, Jensen Huang and Sam Altman took the same San Francisco stage and gave three different answers to the only question that matters this week: who decides how fast frontier AI is allowed…
xAI Watch
A science writer asked two consumer chatbots whether AI is going to end the world, and got the week's argument in miniature. Pressed to argue against its own reassurances, xAI's Grok improvised a gleeful three-phase plan to wipe everyone out. Asked the same question straight, Anthropic's Claude gave…
Jacob Coxon Watch
Jacob Coxon's resignation post, the one warning that AI could kill us all by the end of the decade, has drawn more than 150 million views and carried the argument out of the Silicon Valley bubble and into phone calls home. The sharpest response so far may be…
xAI Watch
xAI's speech-to-speech model has become a shelf product. Grok Voice is now live on fal.ai, the fast-inference platform, with real-time streaming, audio cloning and support for more than 25 languages. The number that should make voice-stack vendors sit up is the price: $0.00083 per second of audio, which…
Yoshua Bengio Watch
About 10 people smashed the windows of Mila, one of the world's best-known AI research institutes, in Montreal late on Tuesday night, and hours later its founder Yoshua Bengio stood in front of Canada's AI establishment and drew his line: the fears driving public anger are legitimate, but…
