NYC Council holds landmark AI oversight hearing, grilling Anthropic, OpenAI, Google, Meta
Executives and whistleblowers testified under oath that AI risks cannot be quantified, as local governments begin taking up AI safety concerns.
Unite.AI · +15 outlets
Nvidia-backed Reflection unveils Beam, first open-weight model aimed at Chinese open rivals
The 501B-parameter mixture-of-experts model is positioned as a Western answer to Chinese open models such as GLM-5.2, at lower compute cost.
MarkTechPost · +13 outlets
Trump stands up 'Super Intelligence Force' AI task force, appoints Jay Clayton to lead it
The new White House task force will coordinate AI policy amid questions about its expertise and role in overseeing the industry.
ktts.com · +14 outlets
OpenAI to watermark ChatGPT text in EU for AI Act compliance, optional for API users
OpenAI will apply invisible watermarks to ChatGPT and Codex text in the EU while letting API customers worldwide opt out, unlike Anthropic.
The Decoder · +13 outlets
Sam Altman says world should accept AI's 'bad things' for its benefits
The OpenAI CEO framed some harm as an acceptable cost of AI progress, remarks that followed a safety leader's resignation.
qz.com · +14 outlets
Meta and Microsoft cut back on Anthropic's Claude in favor of in-house AI
Two of Anthropic's largest enterprise customers are sharply reducing Claude usage as the company shifts from partner to competitor.
The Decoder · +14 outlets
Pentagon stops using Anthropic AI tools after blacklisting company
The Defense Department halted use of Anthropic tools following the company's blacklisting, a blow to its government footprint.
BBC · +14 outlets
Poll finds 77% of Americans want AI development slowed or stopped until safety is verified
A Quinnipiac poll also found 86% support independent safety standards and 74% have little or no trust in AI companies.
The Decoder · +3 outlets
US and China plan AI incident channel as competition tests cooperation
The two governments are discussing a dedicated communications channel for AI incidents amid an intensifying technology rivalry.
voanews.com · +14 outlets
MCP protocol flaw exposes structural risk in agent-to-agent communications
Trust gaps in the widely used MCP protocol let malicious prompts spread between AI agents, raising new liability questions.
Ars Technica · +14 outlets
California finalizes new wave of state AI, privacy and workplace laws
California signed nation-leading chatbot and social media rules and set obligations for employers using AI in the workplace.
JDSupra · +6 outlets
Utah expands health AI pilots, letting startup prescribe acne medication without doctor oversight
Utah is advancing AI pilots for prescriptions and women's health, including a first-of-its-kind acne prescribing service.
Los Angeles Times · +3 outlets
Mamdani calls for AI regulation 'at every single level' as NYC debate intensifies
The NYC mayoral candidate pressed for layered AI rules while rival Menin emerged as a chief antagonist on the issue.
Daily Kos · +13 outlets
Ex-Anthropic researcher and whistleblowers testify as NYC weighs AI curbs
Testimony from a former Anthropic researcher and company whistleblowers added weight to calls for AI restrictions.
Bloomberg Law News · +2 outlets
IRS risk chief issues warning to regulators on AI oversight
The IRS chief risk officer cautioned regulators about the pace at which AI is outpacing existing supervisory frameworks.
BeInCrypto · +3 outlets
Cohere launches North 2 with expanded enterprise AI agent capabilities
Cohere's new North 2 targets enterprises with stronger agent features as competition in the business AI market grows.
HPCwire · +13 outlets
Aleph Alpha releases Kolibri, open-weight model pitched for European AI sovereignty
The German-English 78B-parameter mixture-of-experts model was trained on 768 B200 GPUs and leans heavily on German data.
The Decoder
Federal Reserve-adjacent BIS paper warns AI investment loops may amplify financial risks
The BIS paper flags 'artificial inbreeding' in AI investment as a potential amplifier of financial instability.
INSIGHT EU MONITORING