Anthropic CEO warns of AI risks

The CEO of US artificial intelligence startup Anthropic has warned of the need to slow the pace of AI development.

In an online post on Saturday, Dario Amodei said, "We must slow the pace at which we improve the capabilities of AI models."

He said that because AI is such a powerful technology, its risks are serious. He said they include the risk of losing control of AI systems, misuse of AI for cyberattacks and bioterrorism, and serious economic disruption.

He noted that since this summer, AI has been advancing drastically faster. He warned that left unchecked, it could "outrun our ability to understand and control these systems."

In July, cutting-edge AI models being developed by OpenAI mounted a cyberattack on another company and obtained important information.

Amodei's sense of alarm is shared by other people in the AI sector.

A researcher at Anthropic quit earlier this month, saying "superintelligence," which is being pursued by OpenAI and Anthropic, could threaten human life. The comment made headlines around the world. But some people point out that even if developers choose to slow down, competition with China and other rivals will likely remain increasingly fierce. They say it will not be easy for firms to work together to reduce the pace of AI improvement.