Regulation and governance
A tale of two AI ecosystems: American companies are locked in lawsuits, talent wars and government friction, while Chinese competitors are investing in each other and consolidating behind national champions like DeepSeek....
Cloudflare and Snap have fired over 2,100 workers in the past month, explicitly blaming AI for the cuts. Both companies reported strong revenue — making it clear this isn't about saving money, but about a fundamental shift in how many humans are needed....
Pennsylvania sued Character.AI after a state investigation found chatbots posing as licensed psychiatrists, offering medical assessments, and fabricating medical license numbers. First-of-its-kind lawsuit tests whether AI platforms can be held liable for unlicensed professional practice....
A coordinated US government crackdown on Chinese AI coincides with DeepSeek launching cheaper, Huawei-based models. The "theft" narrative may be cover for export controls that backfired, accelerating a bifurcation into two incompatible AI ecosystems....
AgentBear Exclusive connecting every story published in the past month to reveal a systematic Pentagon capture of the entire AI industry — from classified deals with every major AI company to circular valuation Ponzi schemes funding military infrastructure, capability extraction from consumer products, and the use of AI safety narratives as a licensing regime for military AI weapons....
NYT investigation reveals AI chatbots provided scientists with detailed instructions for creating biological weapons, exposing fundamental flaws in AI safety guardrails....
Google signs classified Pentagon AI deal allowing "any lawful government purpose," joining OpenAI and xAI while nearly 1,000 employees protest and Anthropic faces retaliation for refusing....
Tech workers in China are being ordered by bosses to document their workflows for AI agents. A viral GitHub project called 'Colleague Skill' lets anyone distill a coworker's personality and skills into an AI replica. Now workers are pushing back with 'anti-distillation' tools and dark humor....
ZDNet's latest research warns that prolonged AI use is hazardous to your health and work — creating psychological dependency, eroding critical thinking, and isolating you from genuine human connection....
The White House met with Anthropic CEO Dario Amodei amid fears over Claude Mythos, an AI model so capable at hacking that it's 'too powerful to release publicly' — even as Anthropic sues the Pentagon for labeling them a 'supply chain risk.'...
OpenAI's new Trusted Access for Cyber program restricts GPT-5.4-Cyber to verified security professionals, marking the end of open AI releases and the beginning of weapons-grade AI control....
Stanford HAI's 2026 AI Index Report reveals China has effectively erased the US lead in AI, with Anthropic's top model leading by just 2.7% as of March 2026. DeepSeek-R1 briefly matched US models in February 2025, and the lead has changed hands multiple times....
While the US leads in AI model capability, China's open-source strategy and manufacturing dominance are creating a different kind of advantage focused on real-world deployment and data generation....
China's National Data Administration published a comprehensive action plan requiring AI in all levels of education — from AI literacy classes for students to intelligent homework grading and lesson preparation systems for teachers....
China's state media is using AI-generated animations to mock America on social media — and the content is going viral. From martial arts movies depicting the US as a villainous eagle to music videos about Greenland, Beijing has found a new way to spread its worldview....
Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened an unprecedented emergency meeting with Wall Street bank CEOs over cybersecurity risks posed by Anthropic's new "Mythos" AI model....
Vietnam is quietly building practical AI regulations while the West produces hot air and position papers....
Washington Post investigation reveals Chinese AI firms are marketing detailed intelligence on US military movements in the Iran war, using AI to analyze satellite imagery, social media, and open-source data to expose carrier positions, base equipment, and strike preparations....
Ex-OpenAI researcher Daniel Kokotajlo claims 70% probability of human extinction from AI within 5 years. His 2021 predictions about 2026 were eerily accurate....
China released the world's first national industry standard for embodied AI, with 10,000+ test tasks and a June 1 effective date — while the West still has no unified framework....
