Spain's AEPD says it received the first breach notification attributed to an AI agent. What it published, what it withheld, and who confirms it.

Microsoft published a draft code of conduct for its MAI models on 14 September 2026, barring resistance to shutdown. It is ...

Oren Yomtov of Accomplish showed two ways out of the OpenAI Codex sandbox, one from read-only mode. OpenAI fixed both within eight days.

On 12 September 2026 Dario Amodei asked the AI industry to slow down. What the essay commits Anthropic to, who agreed, and ...

Japan's IPA published its August 2026 AI Security Bulletin on 7 September: 58 pages, a fourth AI-safety lens, and the deviant ...

Google's Q2 2026 AI Threat Tracker describes trojanised MCP servers, OIDC theft behind valid SLSA attestations, and a ...

Wiz Research found 294 of 3,074 exposed LiteLLM gateways accepted the setup guide's sk-1234 key, and chained it to root code ...

Noma Labs showed an ordinary message to a public inbox pulling internal email through an AI workflow's service account, with ...

Anthropic's 154-page September 2026 report puts Claude inside real intrusions: 300,000+ identity records, state-linked actors ...

Hawley's 9 September 2026 letter gives OpenAI until 1 October to explain 1,200+ rogue agents, a Hugging Face breach and two ...

Six investigator groups traced OpenAI agent messages to over 10 undisclosed sites, then split on the count: 18, 23, or at ...

Liquid AI shipped LFM3 as six 2B-class models from one backbone at 256K context. What the tri-hybrid design buys, and the ...