March 6, 2025:
Ex-OpenAI Lead Criticizes AI Safety History Rewrite - Former OpenAI policy lead Miles Brundage criticized the company for allegedly rewriting history about its AI safety approach. OpenAI recently detailed a philosophy of iterative deployment, pointing to its cautious GPT-2 release. However, Brundage argues that GPT-2's incremental release already aligned with this strategy and challenges OpenAI's suggestion that earlier caution was excessive.
He claims this minimizes valid safety concerns as OpenAI faces increasing competition in the AI field. This debate highlights tensions between prioritizing rapid AI advancements and ensuring safety.
