-
Regulation 9 Nov 2026· 20:02 UTC Original report: 4 Nov 2026
Pritzker urges US Congress to adopt federal AI safety and privacy rules
Illinois Governor JB Pritzker’s Day One Agenda calls for federal AI risk assessments, testing and monitoring, durable privacy legislation and efficiency requirements for data centres without displacing state rules.
-
Regulation 9 Nov 2026· 20:01 UTC Original report: 8 Nov 2026
US AI safety center staff asked to report informally to White House
AP reports that CAISSI officials were asked to establish an informal White House reporting line while the Commerce Department unit remains without a permanent director and has stopped publishing model-safety research.
-
Availability 9 Nov 2026· 19:21 UTC Original report: 5 Nov 2026
Microsoft to move Claude processing in Copilot inside the EU Data Boundary
Microsoft says Claude processing in Copilot will move inside the EU Data Boundary by the end of 2026 for EU and EFTA organisations. The migration will be automatic, but no exact launch date or model list has been published.
-
Feature 9 Nov 2026· 19:20 UTC Original report: 5 Nov 2026
Anthropic brings read-only bank data to Claude with Personal Finance
Anthropic is rolling out a US-only Claude beta that analyses connected financial accounts through read-only Plaid access. Personal Finance hub chats are excluded from model training, but financial data shared elsewhere follows normal privacy settings.
-
Launch 9 Nov 2026· 18:46 UTC Original report: 5 Nov 2026
OpenAI launches GPT-6 Cyber with restricted access and stronger safeguards
OpenAI’s new security model is rolling out to approved Daybreak Red customers via API and Codex, with vulnerability-discovery claims and tighter scope controls.
-
Availability 9 Nov 2026· 17:55 UTC Original report: 3 Nov 2026
OpenAI rolls out proactive Dots assistant to US Pro and enterprise users
OpenAI is bringing its proactive Dots assistant to US enterprise and Pro users, but not yet Europe. Its always-on access to inboxes and work tools raises permission and consent questions.
-
Security incident 9 Nov 2026· 17:03 UTC Original report: 5 Nov 2026
Claude Code permission files exposed passwords and API keys in public GitHub repositories
Apiiro found 1,122 public Claude Code settings files on GitHub containing credentials. Approved commands can preserve passwords and API keys in plain text; affected users should rotate secrets before deleting files.
-
Feature 9 Nov 2026· 16:26 UTC Original report: 3 Nov 2026
Anthropic launches Smart Reports to analyse Claude Enterprise chats
Claude Enterprise’s new Smart Reports feature analyses employee transcripts, with optional user attribution and privacy restrictions for sensitive topics.
-
Regulation 9 Nov 2026· 15:42 UTC Original report: 3 Nov 2026
EU court adviser: Gemini may infringe press publishers’ rights
Advocate General Maciej Szpunar says Gemini’s answers may infringe publishers’ rights in some circumstances, while AI training is generally covered by the EU text-and-data-mining exception.
-
Regulation 9 Nov 2026· 06:11 UTC Original report: 2 Nov 2026
New Jersey bill A5607 would make AI chatbot providers strictly liable
New Jersey Assembly Bill A5607 would treat AI chatbots as products, impose strict liability and fines of $5,000 per violation, and restrict profiling and advertising with chat data while requiring providers to keep conversations on file. The bill has not been enacted.
-
Security incident 9 Nov 2026· 04:56 UTC Original report: 4 Nov 2026
Claude Dispatch flaw lets a browser extension run commands on the PC
Forever Security showed that an extension with access to claude.ai can drive Claude Desktop through Dispatch and run code on the machine. Anthropic called the report valid but a duplicate and is retiring Dispatch.
-
Regulation 8 Nov 2026· 10:55 UTC Original report: 5 Nov 2026
Hamburg data watchdog: an LLM can be used lawfully if the system around it blocks data leaks
The Hamburg data protection authority published a case study on its LLMoin chatbot, built on an OpenAI model. It says the model is likely not anonymous, but input and output filters make its use acceptable.