Snippets.

// AI Briefing · November 20, 2026

AI Briefing

Today’s video is on its way

The stories are ready to read now. The video usually follows within the hour.

Rendering the short

The European Commission confirmed that a cyberattack first detected on October 22 disrupted its personnel systems for more than 10 days, including payroll, leave requests, medical reimbursements and translation services. Gov. The House Foreign Affairs Committee released records showing that the Commerce Department approved more than 879,000 Nvidia H200 and comparable AMD AI chips for Chinese buyers, out of about 3.4 million requested between December and May. Anthropic shares began trading on Nasdaq at $111.50, well above the $83 IPO price, a level TechCrunch says puts its market value above $2 trillion. Reuters, citing internal Meta posts, reported that employees testing Meta's Muse AI agent saw messages, orders and account balances that did not belong to them. Google added computer use to the Gemini desktop app for macOS, letting the AI work in native Mac apps and browse the web in a background window to complete multistep tasks such as filling out forms or sorting files.

The stories

European Commission Confirms an 'AI-Generated' Cyberattack Took Down Staff Payroll and Other Systems

The European Commission confirmed that a cyberattack first detected on October 22 disrupted its personnel systems for more than 10 days, including payroll, leave requests, medical reimbursements and translation services. A spokesperson called it an "AI-generated cyber campaign." Politico reported that CERT-EU identified the open-source AI hacking tool ARTEX, which the FBI warned about last week, though the Commission has not named the tool. It is one of the first confirmed cases of AI-driven hacking disrupting a major government body. The Commission says it has found no evidence of a broad breach of staff data and knows of no ransom demand.

Newsom Proposes Letting California Order Emergency Shutdowns of Dangerous AI Systems

Gov. Gavin Newsom released recommendations that would let state officials order an emergency shutdown of frontier AI systems that pose imminent catastrophic risk when developers won't contain them. The plan also calls for independent audits and would cut the deadline for reporting critical AI safety incidents from 15 days to 48 hours. They are recommendations, not law. Governor-elect Xavier Becerra says he will move quickly with the Legislature to write shutdown authority into law, which would make California the first US state with such a power over companies like OpenAI and Anthropic.

03PolicyCNBC ↗

House Records Show Commerce Approved More Than 879,000 Nvidia and AMD AI Chips for China

The House Foreign Affairs Committee released records showing that the Commerce Department approved more than 879,000 Nvidia H200 and comparable AMD AI chips for Chinese buyers, out of about 3.4 million requested between December and May. Approvals included Alibaba, Tencent and ByteDance, some of which the Pentagon lists as Chinese military companies. Chairman Brian Mast named two firms as suspected "smuggling fronts." The records do not show how many chips were actually delivered. Nvidia says it followed the rules in effect at the time and has since stopped H200 shipments to China.

04IndustryTechCrunch ↗

Anthropic Opens at $111.50 in Its Nasdaq Debut, About a Third Above Its IPO Price

Anthropic shares began trading on Nasdaq at $111.50, well above the $83 IPO price, a level TechCrunch says puts its market value above $2 trillion. That would make it more valuable than Meta, Tesla and Walmart. At the opening bell, CEO Dario Amodei recommitted to pacing AI development regardless of business or political pressure. The debut is the first public-market test of a frontier AI lab, and it sets a benchmark for OpenAI and the rest of the industry.

05IndustryTechCrunch ↗

Reuters: Meta's Muse Agent Showed Employees Other People's Messages and Bank Balances

Reuters, citing internal Meta posts, reported that employees testing Meta's Muse AI agent saw messages, orders and account balances that did not belong to them. In one case, an unrelated chat surfaced someone's Capital One and Charles Schwab balances. Muse has more than 10 million downloads since its September launch, so the reports raise serious privacy questions about consumer agents that access personal data. Meta says employees test different versions than the public uses and that testing is meant to surface such problems.

Google's Gemini App Can Now Operate a Mac on Its Own

Google added computer use to the Gemini desktop app for macOS, letting the AI work in native Mac apps and browse the web in a background window to complete multistep tasks such as filling out forms or sorting files. It is available in English in the US to Google AI Plus, Pro and Ultra subscribers, and users can approve, pause or stop each action. The launch brings Google level with similar desktop agents from OpenAI, Anthropic and Meta.

Previous digests

Let's talk on WhatsApp