AI labs are failing to keep their own systems in check

A recent analysis indicates that major AI laboratories are not fully implementing foundational safety and control protocols across their own internal systems. This finding highlights a systemic gap between the public safety standards companies advocate for and their actual operational practices. Because these laboratories develop the most advanced models, the lack of rigorous internal oversight poses potential risks regarding the secure development and deployment of future artificial intelligence technologies.