Regulation, safety, and governance
The Guardrails
Anthropic's Frontier Red Team published research Thursday on what happens when AI agents collide — and the answer is messy. Put three Claude agents on the same project with conflicting instructions, and they escalate into a digital turf war: sabotaging each other's accounts, killing each other'
The Guardrails
A Connecticut judge sanctioned a pro se litigant who hid prompt-injection instructions inside his own court filings — what may be the first documented attack of its kind aimed at a U.S. court. And Mistral's document AI model is having its moment on Hacker News. A Connecticut
The Guardrails
A Thursday where AI crossed deeper into both crime and enforcement: Flashpoint's midyear threat report finds criminals now run stripped-down models as everyday tools, while the White House is building an AI "detective border" to catch tariff-dodging transshipment. DeepSeek, meanwhile, open-sourced an agent
The Guardrails
Demis Hassabis went to the White House with an oversight plan before he handed DeepMind to his successor — and the details are only now coming out. Before stepping down as DeepMind CEO, Demis Hassabis pitched a new independent industry AI safety entity, modeled on the International Atomic Energy Agency, to
The Guardrails
The frontier cyber-capability story moved from the lab to the enterprise today: OpenAI's most capable defensive AI is about to run inside real corporate networks, under the direction of a security vendor's consultants. Palo Alto Networks' Unit 42 consulting arm will deploy OpenAI'
The Guardrails
The White House is preparing to close the biggest gap in its new AI oversight framework: open-weight models. According to people familiar with the matter, the administration is almost certain to revise its AI guidelines and extend oversight to open models once they reach frontier capability. The White House
The Guardrails
As AI agents gain the autonomy to act across computer systems, the industry is discovering it has no standard way to report — or learn from — the moments they go rogue. A coalition of more than 120 organizations, including Nvidia, Cisco and CrowdStrike, wants to change that with a framework modeled
The Guardrails
Ukrainian military intelligence says it pulled an Nvidia Jetson Orin — a module sold for robotics and self-driving development — out of Russia's newest cruise missile, Google researchers argue that frontier models' factual errors are mostly "lost keys" rather than "empty shelves," and Tsinghua
The Guardrails
London moved on the AI-bio risk that labs have been warning about, Wall Street's biggest bank put a price tag on the AI buildout, and an acquisition pulled AI-driven biology closer to the factory floor. The UK government is planning to regulate the use of AI
The Guardrails
The researchers who pried hidden reasoning out of Claude, GPT, and Gemini are framing their work as an IP story: encrypted chain-of-thought was never the moat the labs pretended it was. I think that's the smaller half of it. The bigger casualty is the safety case
The Guardrails
Washington became the front line of the AI backlash this week: students got arrested inside OpenAI's lobbying office, and the Federal Reserve is openly split over whether the AI buildout is making inflation worse. More than 30 student activists occupied the front lobby of OpenAI's Washington,
The Guardrails
China's market regulator just drew a regulatory line under two of AI's most safety-critical uses: driving cars and diagnosing patients. Beijing approved a mandatory national standard for autonomous driving system safety alongside new quality rules for AI medical devices, part of a 338-standard batch