AI Safety - Page 15
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
Anthropic Says Claude Proved 239 Open-Source Entry Points Crash-Free
Anthropic and Theorem used Claude to write machine-checked proofs that 239 OSS-Fuzz entry points cannot crash, finding hundreds of bugs along the way.
Anthropic Gives Agent Skills to AAIF, Saying 70+ Platforms Use It
Anthropic has donated Agent Skills to the Agentic AI Foundation, which also adds OpenAI's Open Responses, OpenEnv and OpenSandbox to reach 10 projects.
Suno Admits in Court It Has Not Deleted All AI Training Recordings
Suno's answer to UMG and Sony's second lawsuit admits it still holds some recordings used to train its AI, as the labels claim v6 'launders' them.
Reservoir Sues Anthropic and Suno, Alleging AI Trained on Its Songs
Reservoir has sued Anthropic and Suno in California, alleging they copied hundreds of its songs to train AI. It seeks up to $337.8m combined.
Anthropic’s New Claude API Version Rejects Temperature Settings
Anthropic's Claude API version 2026-10-21 returns 400 errors for temperature, top_p and top_k. New SDKs send it by default and move to Apache 2.0.
South Korean Bill Would Make Firms Report AI-Driven Mass Layoffs
A South Korean lawmaker has proposed making firms report layoffs caused by AI, with quarterly official statistics. It is a bill, not law, JoongAng says.
Anthropic Seeks Fair-Use Ruling on Claude Lyrics as Judge Weighs First AI Trial
Anthropic asked a US judge to rule that training Claude on song lyrics is fair use. The judge did not rule, but weighed whether it should go to trial.
Palantir Says Neither It Nor the US Can Switch Off NATO’s Maven AI System
Palantir's Europe chief told Reuters its Maven AI used by NATO has "no kill switches", as Starlink disruptions revive European fears of relying on US tech.
Anthropic’s Claude Duet Agent Will Not Move Money Even With Approval
Anthropic has detailed Claude Duet's safeguards: approval cards, a two-tier privacy classifier, passwords the model never sees, and actions always barred.
Anthropic’s Draft IPO Filing Shows $16.3bn Half-Year Revenue, WSJ Reports
Anthropic made $16.3bn in first-half revenue and lost $80.7bn, mostly non-cash, a draft IPO prospectus seen by the WSJ shows. It is not public.
Latest on TSN
Explore the TSN Atlas: Every topic we cover, mapped. Click a bubble to dive in.