Anthropic CEO admits tech industry misled public about AI risks

Anthropic CEO and co-founder Dario Amodei acknowledged that the technology industry has misled the public about the dangers of artificial intelligence, warning that the technology poses real risks that could threaten humanity if urgent measures are not taken.

“I’m not going to lie to you: there are real dangers. And I think for too long the industry has lied to people about the risks of this technology,” Amodei said in an interview with CBS News’ Sunday Morning.

The executive also acknowledged that he had not fully understood what it would look like for AI to advance as rapidly as it has, describing its development as an exponential curve that is becoming increasingly steep.

Amodei said the situation does not mean people should panic or shut down AI entirely, but called it “a warning sign” that the industry needs to slow down. His comments come after former Anthropic and OpenAI researcher Jacob Coxon warned that the companies are “betting with our lives” by developing self-improving superintelligence.

In response to the concerns, Amodei said Anthropic will give independent evaluators permanent access to its models. He also backed federal regulation, including the possibility of a mandatory “off switch” capable of shutting down dangerous AI systems.

However, Amodei rejected a total ban on superintelligence, arguing that giving up its potential benefits, including possible medical cures, would not be the right approach.

The Anthropic CEO also said he feels “uncomfortable” with such powerful technology being developed by private companies. He suggested that AI governance could involve “a combination of democratically elected governments.”

“If we build it the right way, I think the probability of something bad happening is very low. If we build it the wrong way, the probability of something bad happening is very high,” Amodei concluded.