‘AI tsunami’ is coming, world dangerously underprepared: Anthropic CEO Dario Amodei

Update: 2026-02-24 18:30 IST

New Delhi:  Anthropic CEO Dario Amodei has warned that an “AI tsunami” is approaching and said artificial intelligence is reaching human-level capabilities much faster than most people realise.

Speaking on a podcast with Zerodha co-founder Nikhil Kamath, Amodei said society is not fully prepared for the economic and geopolitical changes that advanced AI systems could bring.

He compared the current situation to a tsunami visible on the horizon, saying people can see it coming but are still convincing themselves that it is not real.

“It’s as if this tsunami is coming at us. It’s so close, we can see it on the horizon, and yet people are coming up with explanations like, ‘Oh, it’s not actually a tsunami, it’s just a trick of the light,’” he said during the discussion.

Amodei said AI’s impact on industries will vary depending on the nature of tasks involved.

He explained that human-centered roles, which require deep interaction and emotional understanding, may take longer to be disrupted.

“However, tasks such as coding, mathematics and scientific research are increasingly being handled by AI systems,” Amodei mentioned.

The discussion also addressed evolving assumptions around data as the primary driver of AI progress.

“When you train on math or coding environments, you’re not really getting data… it’s more synthetic. You’re creating the data,” he said, adding that “dynamic data that the model creates itself… is becoming more important.”

Reflecting on AI’s broader trajectory, Amodei reiterated a balanced perspective that acknowledges both transformative potential and associated risks.

“My instinct is we’re about to cure a lot of diseases,” he said, while also noting the importance of responsible development and governance.

He noted that while AI models are already doing a growing share of coding work, the broader responsibilities of software engineering will take more time to automate.

Still, he believes that end-to-end automation of such tasks could happen sooner than many expect.


Tags:    

Similar News