METR chief warns AI safety research hampered by severe talent shortage
Beth Barnes says METR struggles to find enough researchers despite high salaries, limiting its ability to keep pace with rapidly advancing AI models.
Beth Barnes, who left OpenAI to launch METR in 2022, warns that the nonprofit’s ability to monitor AI progress is throttled by a lack of skilled researchers, despite posting salaries exceeding $500,000. METR works with OpenAI, Anthropic, Google and Meta to deliver independent evaluations, but competition for talent remains fierce. The lab’s small staff of 35 has been stretched thin, especially after OpenAI’s recent hack involving Hugging Face, which underscored the need for external safety audits.
METR’s widely cited capability chart shows AI abilities doubling roughly every seven months, and its team has already flagged potential rogue deployments in unreleased models. President Chris Painter sees METR as a “humanity’s preparedness team” and suggests that forthcoming regulations requiring external safety audits could boost hiring. Researchers like Neev Parikh and Ajeya Cotra remain optimistic that growing public and legislative focus will eventually expand the field.
Why it matters
A shortage of AI safety experts could slow oversight of powerful models, increasing risks to society.
In this story
Related stories
4 in this thread