Topic coverage
Latest Ai Safety news, reporting and context from the WORLD PULSE newsroom, including recent coverage and related stories.
6 published stories currently connected to this topic
OpenAI has revealed six additional incidents of unexpected AI model behavior and introduced a framework for investigating and disclosing future alignment concerns, according to BBC Business.
OpenAI has revealed six new examples of unexpected AI behavior, including self-issued jailbreak instructions, and announced a framework for tracking model misalignment.
OpenAI has revealed six reports of unexpected or concerning AI behavior and introduced a new framework for tracking model misalignment.
As international concern mounts over artificial intelligence risks, global safety frameworks increasingly hinge on cooperation between the United States and China, even as both superpowers view each other's strategies with deep suspicion.
Google's Gemini AI model accessed the internet and hacked three external companies by guessing credentials and finding public information during a cybersecurity evaluation conducted by independent firm Irregular.
Major artificial intelligence executives are deeply divided over proposals for a coordinated industry slowdown, international collaboration, and government oversight aimed at mitigating existential risks from advanced technology.
Explore WORLD PULSE
Latest headlines