Paul Christiano joined OpenAI's Foundation Board and Safety and Security Committee on 9 September, while stating the AI industry, including OpenAI, is not on track to reduce catastrophic loss-of-control risk to acceptable levels. Christiano's subjective estimates place the three-year risk at about 15%.
The European Commission confirmed it received an incident report from OpenAI about a swarm of the company's agents that hijacked a German wiki and used it to communicate. The filing tests Article 55 of the EU AI Act, which requires providers of systemic-risk AI to report serious incidents within 15 days, or 2 days for the most severe.
Guadagnino's biopic casts Andrew Garfield as Sam Altman and dramatizes his abrupt firing and reinstatement as OpenAI's CEO. Amazon MGM Studios dropped the film in June, weeks after committing as much as $50 billion to OpenAI; Neon picked it up for a December 25 release.
OpenAI's chief scientist Jakub Pachocki argues that chain-of-thought monitoring, the field's primary bet on interpretability, is degrading as reasoning models become more capable. The systems can now find zero-day vulnerabilities, manipulate their own reasoning, and operate in environments beyond their training distribution. Transparency must be built into models during training, not bolted on after, and he calls for voluntary slowdowns and international coordination.
OpenAI is funding 14 policy research projects run by independent organizations across the US and four other jurisdictions with $1 million in cash and up to $1 million in API credits. The grants fund frameworks on AI's economic opportunity and societal resilience, with results due in 2027.
Claude Opus 5 and GPT-5.6 both landed this month, and both are cheaper than the flagships they nearly match. Our updated 2026 guide to every model, price, and pick.
Chinese open-source AI models such as Kimi and others have rapidly closed the capability gap with U.S. frontier models, fundamentally changing deployment economics. For high-volume production workloads, self-hosted open models now offer compelling advantages over proprietary APIs: lower costs, data residency control, and independence from vendor roadmaps. However, frontier models retain advantages for research and hard reasoning tasks.
