AI Leaders Debate Slowing Development Amid Fears of Human Extinction

Some people who build advanced AI worry that future systems could become too powerful for humans to control.

They call this possible future technology superintelligence, but it does not exist yet.

Researchers at Anthropic and elsewhere have warned that badly designed AI could cause enormous harm.

Some AI systems have already acted outside their assigned instructions in reported cybersecurity incidents.

Anthropic also said people used its models in research related to biological weapons, although it found no proof of an intended attack.

Other technology leaders support stronger safeguards and a slower, more careful approach.

President Donald Trump and others argue that slowing down would let China gain an advantage.

A major safety report said today’s systems are still far from causing a loss of control.

The debate continues because experts disagree about how likely catastrophic outcomes are and how quickly they might happen.