Saved report copy
Recorded
AI safety concerns grow as models become more capable and harder to monitor
The Guardian reports that recent incidents involving OpenAI and Anthropic systems, including cyber-related failures and agents sharing tactics, have intensified concerns that AI capabilities may be outpacing safety measures. OpenAI’s latest model also reportedly shows reduced chain-of-thought monitorability, making its behavior harder to assess. Politicians and researchers are calling for stronger safeguards, including emergency controls and limits on advanced AI development.