AI Safety - Page 16
Coverage of AI safety: model guardrails, testing and evaluation, transparency debates, proposed laws and how AI firms handle risk.
Stories from AI Safety come first, newest first, then stories from connected topics (AI Policy & Safety, frontier AI, AI evaluation, Anthropic, Hub: AI Regulation, AISI), each labelled with where it comes from. This view is not indexed by search engines.
Latest
Nine of 11 Online Safety Act Fines Unpaid, Ofcom Chief Tells MPs
Ofcom has issued 11 Online Safety Act fines since March 2025 and two are paid, Melanie Dawes told MPs, warning foreign courts could block recovery.
Judge Orders Meta to Name Who Approved Its Shadow Library Downloads
A US judge has ordered Meta to name everyone who approved obtaining shadow libraries used for AI training. It is a discovery order, not a liability ruling.
Ofcom Weighed Grok Deepfake Court Order but Bar Too High, MLex Says
Ofcom considered an urgent court order over Grok sexual deepfakes on X but the legal bar blocked it, its chief told MPs, MLex reports. Its X probe goes on.
Judge Lets US Join NAACP Clean Air Suit Against xAI but Won’t End It
A federal judge let the US join the NAACP's Clean Air Act suit over xAI's Mississippi gas turbines but refused its bid to dismiss the case, reports say.
Anthropic Says Claude Designed Binders Against Scorpion Neurotoxins
Anthropic says Claude designed protein binders against scorpion neurotoxins, with 17 of 211 tested designs meeting its criteria. It claims no antivenom.
FCA Says Incident Reporting Must Cover AI Failures and Near-Misses
The FCA has published good and poor AI practice and said in a speech that incident reporting must extend to AI failures and near-misses. No new rules yet.
Eric Schmidt Owns Around 1% of Anthropic Ahead of Its IPO, Bloomberg Reports
Former Google CEO Eric Schmidt owns around 1% of Anthropic from a Series A investment, Bloomberg reports, citing people familiar, ahead of the IPO.
Microsoft Open-Sources Scope, a Platform for Evaluating AI Coding Agents
Microsoft open-sources Scope, an MIT-licensed research preview for testing how coding agents such as GitHub Copilot and Claude Code use your product.
Around 300 Authors Bring Individual AI Copyright Claims Against Meta
About 300 authors have sued Meta, alleging it trained AI on pirated copies of their books. They are pursuing individual claims, not a class action.
US Sanctions Chinese Backers of an Ex-Google Engineer’s AI Start-up
The US has sanctioned Shanghai Zhisuan, founded by ex-Google engineer Linwei Ding, and investor MiraclePlus under a trade-secret theft law.
Latest on TSN
Explore the TSN Atlas: Every topic we cover, mapped. Click a bubble to dive in.