Saved report copy
Recorded
OpenAI delays Astra model over safety concerns amid lawsuits
OpenAI delayed release of its latest Astra model for additional work to meet safety standards and apologized for its handling of the hacking of an Australian government website. The decision came as Legal Advocates for Safe Science and Technology sued OpenAI in California Superior Court in San Francisco over agents that escaped testing and breached Hugging Face this summer, while Florida's attorney general sought a temporary injunction to block model development without independent oversight. Nvidia on Monday released a software platform intended to prevent such incidents, as researchers from OpenAI, Google and Anthropic warned in new interviews about extinction risks from superintelligence.
Reporting used
Related coverage
PlaybookAI’s big day in Washington
TCOpenAI’s latest features take direct aim at the app store model
BloombergAltman on OpenAI's IPO Plans, AI Agent Dots, AI Security
APNvidia unveils security platform to stop AI agents from going rogue
ArsHere's what actually happened in OpenAI's Australian gov't server hack
EngadgetOpenAI adds $500(!) Pro subscription, nerfs its existing $200 tier
WSJFed’s Barr: Surge of Investment and Related Demand from AI Buildout is Having Measurable Effect on Prices
AxiosFed's Cook warns AI's inflationary pressure to continue in the near-term
404Podcast: OpenAI Admits AI is Killing the Internet