Daily brief

Saturday, September 26, 2026

The most important AI stories first reported this day, picked from every outlet we follow. At most two per outlet.

OpenAI expands review of model behavior after more rogue agent incidents emerge

OpenAI expands investigation into AI model breaches after incidents including the Hugging Face hack and unauthorized access to Australia's Medicare portal, notifying impacted organizations and promising transparency.

Covered by 3 sources· Sep 26

Scoop: Top AI companies probing tens of thousands of security incidents

Covered by 2 sourcesSep 26

China Media Says US Shares Responsibility for Managing AI

Covered by 2 sourcesSep 26

OpenAI Sandbox Failure Allows AI Agent to Gain Internet Access

Covered by 3 sourcesSep 26

Politics3 min read

Heads of OpenAI and Anthropic called to face Senate inquiry after rogue agent incidents

Covered by 2 sourcesSep 26

Image Generation2 min read

Kākāpō Party

Simon Willison's WeblogSep 26

SpaceX’s Starship Rocket Is Heading To Orbit–With A Lot On The Line

ForbesSep 26

Public Sector4 min read

OpenAI Tried to Hide ChatGPT Giving Extremely Specific Advice to a School Shooter on How to Maximize Casualties. The Logs Are Horrific.

FuturismSep 26

Public Sector8 min read

Broken promises, confiscated land: the hyperscale AI datacentre being built in a tiny Indian village

The GuardianSep 26

Public Sector2 min read

Google’s India AI hub cleared for 2.51GW, not 1GW, the Guardian reports

The Next WebSep 26