AI needs a 'brake pedal', warns Anthropic co-founder

Article excerpt

Anthropic co-founder Jack Clark warned BBC's Newsnight that artificial intelligence could advance to a point where it develops independently of human oversight, and suggested the technology needs safeguards analogous to a 'brake pedal.' Clark's concern centers on AI systems reaching levels of sophistication that outpace human ability to monitor or control them. The warning comes as major AI developers face mounting pressure from regulators and safety advocates to implement more rigorous oversight mechanisms. Anthropic, which Clark helped establish, has positioned itself as a leader in developing AI systems with built-in safety measures.