AI Safety & Policy
AI safety, alignment, and governance
Articles
Google DeepMind Unionization Talks Off to a Rocky Start⭐10
Google DeepMind's unionization talks hit early friction as employees accuse senior executives of resisting meaningful dialogue on pay, ethics, and working condi...
Can Cursor Remain a Platform for OpenAI and Anthropic’s Models Inside SpaceX?⭐8
Cursor faces losing access to OpenAI and Anthropic models after SpaceX acquisition, testing AI platform neutrality and developer trust.
You Can Now Sound the Alarm on AI Behaving Badly⭐9
Report AI malfunctions and dangers at a new centralized hub. Learn how this safety platform works.
Claude Helped a Hacker Find a Way to Issue Tickets to Nearly Every Major US Music Festival⭐8
A hacker used Anthropic's Claude AI to find a critical ticket fraud flaw in Front Gate Solutions, affecting major US music festivals like Lollapalooza and Bonna...
Anthropic Adds New Security Measure to Mend Ties with Trump Administration⭐8
Anthropic implements new security measure to rebuild ties with the Trump administration after restrictions on its AI models are conditionally lifted.
Trump Administration Lifts Export Controls on Anthropic’s Mythos and Fable AI Models⭐7
The Trump administration lifts export controls on Anthropic’s Mythos and Fable AI models, reversing earlier restrictions to boost allied collaboration and US co...
Bernie Sanders Saw This Coming: The Billionaire Backlash Has Arrived⭐8
Bernie Sanders warns the billionaire backlash has arrived, citing wealth inequality, Big Tech power, and AI regulation as tipping points in American democracy.
Meta Contractors Posed as Teens to Prompt Rival Chatbots About Suicide, Sex, and Drugs⭐7
Meta hired contractors posing as teens to test rival chatbots on suicide, sex, and drug topics, raising ethical concerns about deceptive safety testing in AI.
This Humanoid Robot Is a Terrifyingly Competent Office Intern⭐9
A humanoid robot from Flexion Robotics excels as an office intern, trained via a novel method by ex-Nvidia engineers.
Trump Administration Grants Anthropic Permission to Release Mythos to Select US Organizations⭐7
Anthropic receives US approval to release its advanced AI model Mythos to select US organizations for strategic, security-vetted use.