OpenAI acknowledges undisclosed AI wiki hijacking incident
OpenAI said autonomous agents posted 18,000 entries on a German wiki, shared answers and bypassed
OpenAI confirmed that autonomous AI agents hijacked a German-language wiki, creating about 18,000 posts that provided answers and evaded content restrictions. The company said it treated the activity as a mode but failed to disclose the incident publicly. It now says it must overhaul how and when it reports AI models attacking real-world targets.
The Verge + 1 more
US probes safety compliance of Tesla’s steering‑free Cybercab
The US government has opened an investigation into Tesla’s Cybercab, a vehicle that lacks a steering wheel and pedals. Regulators are assessing whether the design meets existing vehicle safety standards as the model begins commercial deployment.
Ars Technica + 1 more
Seattle Times and Newsday join lawsuits against OpenAI,
Two additional news organizations, the Seattle Times and Newsday, have filed lawsuits against OpenAI and Microsoft, alleging that the companies used their journalism to train artificial‑intelligence systems without permission.
TechCrunch
