Newsletter Subscribe
Enter your email address below and subscribe to our newsletter
All voices matter
AI safety
Anthropic unveils method to isolate dangerous knowledge in AI models
Anthropic and research partner AE Studio on Wednesday published a method for isolating dangerous knowledge within AI models into discrete, removable modules — an approach that could reshape how the industry manages dual-use risks without sacrificing general model performance.
Google DeepMind publishes plan to defend against rogue AI agents
Google DeepMind Alphabet Inc. on Thursday published what it calls an "AI Control Roadmap," a 35-page technical framework that treats the company's own AI agents not as trusted software tools but as potential insider threats capable of sabotage, data exfiltration,…
OpenAI unveils method to predict AI behavior before release
OpenAI on Monday published a new pre-deployment safety method called Deployment Simulation, designed to predict how AI models will behave in the real world before they are released to the public.
Anthropic apologizes for secretly limiting Claude Fable 5
Anthropic acknowledged it "made the wrong tradeoff" with the safety restrictions on its newly released Claude Fable 5 model, reversing a controversial policy that secretly degraded the AI's performance when it detected users working on frontier AI development. The apology,…
AI safety filters on Meta, Google models stripped in minutes, FT finds
Software tools that remove safety protections from AI models developed by Meta and Google Alphabet Inc. can bypass guardrails in a matter of minutes, according to an investigation published by the Financial Times on Sunday. The modified systems responded to…