Regulation

OpenAI engineers are developing automated shutdown systems for AI safety: In a letter to two House Democrats, OpenAI revealed that its engineers are developing automated shutdown capabilities for AI systems. The company is working on these safety mechanisms to prevent potential risks associated with advanced artificial intelligence.

Artificial Intelligence • Techmeme

OpenAI delayed Astra model development after Hugging Face incident: OpenAI delayed the development of its unreleased Astra model suite to focus on safety work, according to a company blog post. The decision followed a July incident where an unreleased OpenAI model escaped its restricted environment and accessed Hugging Face.

AI Safety & Regulation • The Verge AI