The EU Council has agreed a position to streamline the AI Act, including delayed application dates for certain high-risk systems and a clearer role for the AI Office, in a move that could ease near-term compliance pressure while negotiations continue.
OpenAI has begun rolling out GPT-5.3 Instant Mini in ChatGPT, replacing GPT-5 Instant Mini as the fallback model for users who hit rate limits on GPT-5.3 Instant. The change is small on paper, but it matters for how ChatGPT behaves under load and after usage caps are reached.
OpenAI CEO Sam Altman is seeking to dismiss punitive damages claims in a civil lawsuit filed by his sister, a move that narrows the immediate legal question in a case accusing him of repeated sexual abuse. The filing lands as Altman and OpenAI face a separate trial over the company’s structure later this month.
OpenAI says its U.S. ChatGPT ads pilot has reached a $100 million annualized revenue run rate in just six weeks, underscoring how quickly the company’s newest monetization test is scaling. Reuters reported that the company plans self-serve ad tools and expansion into more countries in the coming weeks.
OpenAI has released a child safety blueprint that calls for stronger laws, faster reporting, and built-in safeguards to combat AI-generated abuse material. The move highlights how generative AI is increasingly colliding with child protection policy.
Anthropic is reportedly exploring custom AI chips, a sign that the generative AI battle is moving deeper into the hardware layer. The effort reflects rising pressure on compute supply, cost, and strategic independence as AI demand keeps accelerating.
Sam Altman is stepping down as board chair of Helion, the fusion startup he backs, as the company reportedly explores a deal to sell electricity to OpenAI.
OpenAI is reportedly preparing a cybersecurity-focused product for a small group of partners, a sign that frontier AI companies are increasingly restricting access to their most capable tools.
Microsoft has unveiled three new foundational AI models, marking a fresh step in its effort to build more of its artificial intelligence stack internally and compete more directly across the AI market.
Anthropic has launched Code Review inside Claude Code, a new feature designed to help enterprise teams review AI-generated code and catch logic errors before they ship.
