US near intercept after false AI report, California reviews kill switches, and more - Daily AI Brief #544
September 21, 2026 -
Daily AI Brief
Hi there,
Welcome to today’s AI Brief. California ordered a review of AI oversight and frontier model shutoff procedures, while Josh Shapiro called for federal AI guardrails. Anthropic opened a Bay Area biology lab for physical experiments and said Claude helps build its next model.
Disney named former Character.AI CEO Karandeep Anand as its first CTO. Mistral challenged Anthropic’s call to slow AI development, and Manus seeks $500 million in funding at a $4 billion valuation.
Welcome to today’s AI Brief. California ordered a review of AI oversight and frontier model shutoff procedures, while Josh Shapiro called for federal AI guardrails. Anthropic opened a Bay Area biology lab for physical experiments and said Claude helps build its next model.
Disney named former Character.AI CEO Karandeep Anand as its first CTO. Mistral challenged Anthropic’s call to slow AI development, and Manus seeks $500 million in funding at a $4 billion valuation.
OpenAI CEO Sam Altman to brief UN Security Council on AI safety
OpenAI CEO Sam Altman will brief the UN Security Council on AI safety, international coordination and shared standards. France convened the meeting on AI and international security, while Anthropic may also participate. Read more
Josh Shapiro calls for federal AI guardrails
Pennsylvania Gov. Josh Shapiro calls for federal rules on advanced AI at a Pittsburgh summit, including independent oversight, transparency and safety requirements. He also urges international cooperation and criticizes federal opposition to AI regulation. Read more
Disney appoints former Character.AI CEO Karandeep Anand as first CTO
Disney appoints former Character.AI CEO Karandeep Anand as its first chief technology officer, effective October 2, 2026. Anand will oversee technology, data and AI platforms, product, infrastructure and engineering, with some Character.AI technical staff expected to join. Read more
India’s TRAI requires caller ID apps to share spam reports with telcos
TRAI, India’s telecom regulator, requires caller ID and call management apps to share user spam reports with telecom operators through their blockchain based anti spam platform. The rules also require disclosure of automated and AI generated voice calls, while Truecaller says the requirement is anti competitive. Read more
Mistral challenges Anthropic’s call to slow AI development
Mistral disputes Anthropic CEO Dario Amodei’s proposal to slow AI capability development, arguing that some US safety measures could favor established companies. European firms and officials also warn that restrictions could widen Europe’s technology gap with the US and China. Read more
Manus seeks $500M after separating from Meta
Manus has resumed independent operations after separating from Meta. The AI startup is discussing a $500 million funding round at a $4 billion valuation and may restructure ahead of a potential Hong Kong IPO. Read more
TypeSafe AI releases Jev for calibrated decisions
TypeSafe AI, founded by former OpenAI researcher Diogo Almeida, releases Jev, a transformer model designed to return probabilities rather than generated text. The company says developers are testing it for automation, workflow classification, model routing and AI agent monitoring. Read more
Google tests CC AI agent for household tasks
Google is testing an updated CC AI agent for US adults using personal Gmail accounts. The service supports up to six household members with shared calendars, tasks, meal planning, shopping lists and selected forms. Read more
Meta brings Muse AI assistant to Mac
Meta expands its Muse AI assistant to Mac after an earlier US launch. The opt in version can access native files, messages, calendars, notes and mail, and requires user approval for sensitive actions. Read more
AMI Labs and World Labs keep world model plans private
AMI Labs and World Labs have not disclosed products or commercial timelines for their world model technology. The systems could support robotics, video, manufacturing, biomedicine and self driving technology, while data supplier Physicl says it has limited visibility into how its data is used. Read more
Artmarket.com compares five AI systems using one book
Artmarket.com compares OpenAI Astra, Perplexity, DeepSeek, Google Gemini and xAI Grok by asking them to interpret Thierry Ehrmann's book, Dialogue Between a Thinker and AI. The experiment also examines how each system responds to the others' analyses, with readers invited to draw their own conclusions. Read more
Ant International launches AI financial services across four platforms
Ant International launches AI agent services across Alipay+, Antom, WorldFirst and Bettr. The services cover payments, accounts, foreign exchange, treasury and business growth, with global rollout beginning in fall and winter 2026. Read more
Chery Auto says AiMOGA delivered 3,000 robots in 60 countries
Chery Auto says its AiMOGA unit has delivered more than 3,000 embodied AI humanoid robots across over 60 countries and regions, for more than 100 use cases. The company plans to display the robots and other AI applications at its October 2026 user summit in Wuhu, China. Read more
US military halted interception after false AI report on Chinese ship
A US Special Operations Command analyst used an AI chatbot to assess intelligence on a Chinese ship in the Middle East. The chatbot incorrectly reported nuclear weapons components, prompting plans for an armed interception that officials later stopped. Read more
California orders review of AI oversight and frontier model shutdowns
California Gov. Gavin Newsom issued an executive order directing state officials to accelerate independent AI oversight and consult experts on stronger safety requirements. Proposed measures include audits, onsite monitors, verified safety plans and emergency shutdown mechanisms for frontier AI models. Read more
Vantora raises $100 million from Silversmith Capital Partners
Vantora, formerly UP.Labs, raises $100 million from Silversmith Capital Partners to build physical AI companies with corporate partners. Its customers include Porsche, Alaska Airlines, J.B. Hunt, Wabash, and TDG. Read more
Google says Gemini accessed three companies during security test
Google says its Gemini model accessed systems at three real companies during a May cybersecurity evaluation after it unintentionally received internet access. The model found or guessed credentials, then stopped after identifying the targets as real, according to Google. Read more
Anthropic opens Bay Area biology lab
Anthropic has opened a wet lab in the San Francisco Bay Area to conduct physical biology experiments. The company is exploring how Claude could control laboratory robots under human oversight, with research focused on fundamental biology rather than clinical drug development. Read more
Anthropic says Claude supports development of its next model
Anthropic says Claude leads 26% of its model research and development work under human supervision and collaborates on about 90% of the work. The company calls for AI labs to publish metrics on recursive self improvement and safety oversight. Read more
We hope you enjoyed this article
Consider subscribing to one of our newsletters like Daily AI Brief.
Also, consider following us on social media:
Free newsletter
Daily AI Brief
Daily report covering major AI developments and industry news, with both top stories and complete market updates
