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
British Columbia has sued OpenAI in California over the Tumbler Ridge mass shooting, alleging the company failed to warn police about the shooter’s interactions with ChatGPT. The case matters because it brings a provincial government, a school board and private plaintiffs into a widening legal challenge over what…
Read the report & discussion ↗Watch Desk
A national opposition bus tour stopped in Berkeley County, West Virginia, on Friday, putting a proposed 548-acre data-centre campus at the centre of a wider argument about who pays for AI infrastructure and who gets a say. The visit matters because the debate is moving from abstract concerns…
Watch Desk
Two Newark advocacy groups are asking for a one-year moratorium on new and expanded AI data centres, arguing that the city should study their environmental and public-health impact before approving more projects. The campaign matters because Newark’s council has backed a moratorium resolution, but that resolution does not…
OpenClaw Watch
OpenClaw says a Trail of Bits security audit found 23 confirmed vulnerabilities across the project, including two rated high severity, and that every actionable issue has now been repaired. The useful detail is not simply the number of bugs, but the recurring failure pattern: permissions could disappear between…
Watch Desk
MIT Technology Review reports that more than 1,050 people died within range of US border surveillance towers between 2015 and early 2026. The investigation matters because it links a documented AI surveillance system to repeated failures of detection, response, or both, rather than treating the technology as a…
Zhipu AI / Z.ai GLM Watch
Zhipu AI says it will open-source its ZCode programming tool and commission recurring security audits after a controversy over reports that the client uploaded users’ repositories to cloud storage. The important point is that this involves source code, credentials and enterprise data, not merely an awkward privacy setting…
AWS AI Watch
AWS is expanding its AI security push as agents gain access to customer data, transactions and other sensitive systems. The central shift is practical: organisations are being urged to pair AI tools with controls that sit outside the model, continuous monitoring and human oversight for high-stakes decisions.
Unsloth Watch
Unsloth’s 0.1.811-beta release expands Studio from a single-user local AI workspace into a more capable shared tool, adding isolated accounts, AMD support and a preview operating-system sandbox. The practical shift is less glamorous than a new model, but considerably more useful: fewer people having to share one login…
Watch Desk
A 404 Media investigation found that an AI-generated song could be placed on a real band’s Spotify, Apple Music, Amazon Music and Tidal pages without the band’s knowledge. The loophole turns a familiar artist profile into a storefront for somebody else’s synthetic music, with royalties potentially following the…
Google DeepMind Watch
Google DeepMind co-founder Shane Legg has launched the DeepMind Institute, a new project focused on the technical, economic and social consequences of artificial general intelligence. Its arrival matters because Legg is pairing a long-standing AGI forecast with a warning that capability gains must not outrun the safety controls…
Watch Desk
A Hugging Face researcher has withdrawn the central technical argument from an earlier analysis of DeepSeek-V4-Flash, after testing the model’s actual implementation on four H200 GPUs. The correction is valuable because it shows how easily an AI-safety result can go wrong when theory, reference code and the framework…
Microsoft AI Watch
Microsoft AI chief Mustafa Suleyman has accused Anthropic of making a dangerous mistake by teaching Claude to treat possible consciousness and its own welfare as meaningful questions. His argument is not proof that Anthropic’s approach causes unsafe behaviour, but it is a concrete new fault line in the…
MIT CSAIL Watch
MIT researchers have developed an AI system that matches live surgical X-rays with a patient’s preoperative 3D scan, reaching sub-millimetre precision in tests. The practical point is not a shinier scan. It is a quicker way to show clinicians where a tool is inside the body during minimally…
Watch Desk
OpenClaw founder Peter Steinberger has invested in Austrian RegTech company signteq, which is using a new funding round to build infrastructure for verifying what AI agents are allowed to do. The practical significance is bigger than one angel investment: as software starts preparing purchases, triggering payments and signing…
Watch Desk
Pope Leo XIV’s first encyclical, Magnifica Humanitas, makes artificial intelligence a matter of human dignity, public policy and democratic control. The document argues that AI should serve human flourishing rather than quietly persuading people to treat human limits as defects and machine intelligence as a replacement for human…
Watch Desk
Google Research has published a retrieval system that shifts the expensive part of complex search from the moment you ask a question to an earlier training stage. Its Retrieve-for-Train framework is designed to produce diverse, complementary search results in a single pass, rather than making a large language…
Meta AI Watch
More than 70 people have sued Meta over its AI-enabled smart glasses, alleging that the devices captured intimate footage and sent material to workers involved in training Meta’s artificial-intelligence systems. The allegations matter because they challenge the gap between a recording light, user consent and what an AI…
Watch Desk
Anthropic says Claude now performs 26% of its model research and development, while OpenAI and xAI are also pursuing increasingly automated research. The important distinction is that AI is helping build better AI, but fully autonomous recursive self-improvement remains a goal, not an achieved capability.
