Anthropic's Self-Improving AI Alarms Safety Researchers
A new report from METR, an AI safety research group, says Anthropic is using AI to build more powerful AI. The report says Anthropic's models are improving about twice as fast as they were a year ago. The 11-week investigation found little sign of that slowing down. That worries researchers. They point to the speed of self-improvement and the risk that humans lose oversight. The findings add to the ongoing debate about AI safety and whether strong guardrails are needed.
METR report follows an 11-week investigation into how Anthropic develops its AI
Models are getting better about twice as fast as they were a year ago
No signs of slowing down. Safety researchers are alarmed.
Anthropic is using AI to run parts of its own AI development