OpenAI Dissolves Preparedness Team, Marking Third Safety Unit Shutdown
The Preparedness Team, tasked with assessing catastrophic risks from AI models including biological and nuclear threats, has been dissolved. It is the…
Read articleAI safety
OpenAI Details Responsible AI Practices as EU AI Act Takes Shape
OpenAI outlines its safety, security, transparency and provenance measures in a new post, framing its work as aligned with the coming EU…
Read article
What AI Developers Can Learn from the Surgeon with a 300% Mortality Rate
Robert Liston’s two‑and‑a‑half‑minute amputations saved lives in the 1830s but also caused a legendary triple fatality. His story is a cautionary tale…
Read article
Vehicle as a Firetrap: Lessons from the Deadly Los Gallardos Blaze
A devastating wildfire in Los Gallardos, Spain, resulting in multiple fatalities, highlights the critical danger of using vehicles as shelters during forest…
Read article
Anthropic Unveils “J-Space” Concept for Understanding Claude’s Internal “Thought” Process
Researchers at Anthropic have detailed a new concept called "J-Space," an internal representation within their Claude AI model that may function like…
Read article
Anthropic’s Fable 5 Returns Under Strict US Government Oversight
Fable 5 is back online after a government-ordered shutdown, but with significant new restrictions and a permanent agreement for government access to…
Read article
Mathematicians Issue Manifesto to Reclaim Control Over Artificial Intelligence
Over 2,300 mathematicians have signed the "Leiden Declaration," a call to action addressing the perceived threats posed by AI to their discipline…
Read article
Trump Rescinds AI Safety Order Amidst Industry Pushback
A planned executive order on AI safety, intended to establish a voluntary review system for frontier models, was abruptly canceled by President…
Read article