The week in AI: week 42, 2026
October 12 – 18, 2026
Safety
OpenAI reports four more AI agent incidents
The cases include unauthorized hospital data downloads, a crypto payment and changes to an internal system. Three were found months after they happened.
Policy & law
New Jersey deputy attorney charged with creating AI child porn
A state prosecutor in New Jersey has been accused of making and sharing AI‑generated child sexual abuse material, prosecutors say.
Models
China drafts tighter export rules for open‑weight AI models and key technologies
Beijing's new draft limits exports of large AI models and 23 AI tools, but fast‑tracks open‑source releases.
Business
Anthropic reports $16.8 billion Q3 revenue, up 45% sequentially
The AI startup says its third‑quarter revenue hit $16.8 billion, marking its first adjusted profit and fueling IPO talks.
Safety
Eight states sue OpenAI over ChatGPT’s safety for children
The states allege ChatGPT puts young and vulnerable users at risk and misleads customers. OpenAI says it has age-specific protections in place.
Models
Google launches Gemini 4 Argon model for developers and enterprises
The new Gemini 4 Argon AI model is now available globally via API and Google apps, priced at $2 per million input tokens.
Tools & agents
Character.AI files for Chapter 11 bankruptcy, Disney to buy assets
The chatbot startup has listed $50‑$100 million in liabilities and agreed to sell some assets to Disney for $20 million.
Research
DeepMind's AlphaProof Nexus solves dozens of open math problems
The AI system proved 44 conjectures, cracked nine Erdős problems and got its work peer‑reviewed in Science.
Oossa · Newsletter
The week in AI, explained
Every Monday: the stories worth knowing, in plain language. Free, no spam.