Sign in Subscribe

AI safety

Everything tagged AI safety.

Share your story Subscribe

Latest

News
Anthropic opens Seoul office and signs AI safety MOU with South Korea's MSIT
Anthropic opened a Seoul office on June 17 and signed an AI safety MOU with South Korea's science ministry, days after a U.S. export-control…
Dmytro Spodarets · Jun 20
News
OpenAI introduces Deployment Simulation to predict model behavior before launch
The method replays about 1.3 million de-identified ChatGPT conversations against a candidate model, and caught a reward-hacking behavior OpenAI calls…
Dmytro Spodarets · Jun 17
News
Google DeepMind opens $10M research call on multi-agent AI safety
The grant program, run with Schmidt Sciences, the Cooperative AI Foundation and the UK's ARIA, focuses on risks that emerge when many AI agents…
Dmytro Spodarets · Jun 11
News
Anthropic launches Claude Fable 5 and a cyberdefender-only Mythos 5 model
The pair costs $10 per million input tokens and $50 per million output tokens, with Fable 5 ranking first on two third-party coding and finance…
Dmytro Spodarets · Jun 10
News
Anthropic faces backlash over Claude Fable 5 provision that silently limits researchers' answers
A paragraph in the model's 319-page system card says Fable 5 quietly limits answers about frontier AI development without telling users. Researchers…
Dmytro Spodarets · Jun 10