Major AI Labs Face Growing Pressure to Restrict Computing Power for Safety

OpenAI and Anthropic face mounting calls from lawmakers, researchers, and industry workers to adopt restrictions on computing power for AI development, a proposal aimed at buying time for safety measures but that threatens their competitive edge in the race for more powerful models. Over 1,200 employees from leading AI companies, including OpenAI, Anthropic, and DeepMind, have signed a statement asking the U.S. government to develop technical and governance mechanisms to "deliberately pace the frontier of automated AI development." The push reflects mounting concerns about recursive self-improvement, where AI systems write code for their successors, potentially accelerating progress beyond human comprehension or control.