Amodei Says the AI Backlash Is a Crisis of Trust

Dario Amodei, the chief executive of Anthropic, has a diagnosis for the public's growing wariness of AI: it is not really about AI. Speaking this week, he called the backlash "fundamentally a crisis of trust," and argued that it runs deeper than anything the industry's messaging could fix, as TechCrunch reported.

"Ordinary people don't trust companies, governments, or the tech industry," Amodei said, and they "always suspect that we are cooking up some new way to screw them over," per Fortune. The point was partly self-interested. Amodei has spent years warning about AI's risks, and some critics blame that drumbeat for the sour public mood. He pushed back, saying the suspicion predates both him and the technology.

The numbers behind the mood

The polling supports the shape of his claim, if not every detail. A Pew Research Center survey of more than 5,000 US adults, conducted in February, found that 63 percent think AI is moving too fast and 71 percent believe it makes their personal information less secure. Two thirds had little or no confidence in the government to regulate it well, and roughly six in ten felt the same about the companies building it.

What that captures is not fear of a specific product. It is a low background trust in the institutions doing the building, extended to whatever those institutions ship next. On that reading, an AI lab cannot earn public confidence with a better safety page. It inherits the deficit that banks, platforms and regulators have already run up.

Trust you can check

Amodei's proposed answer, in his telling, is results people can verify rather than promises they have to take on faith. That is easier said than delivered, and it sits awkwardly next to an industry that often asks for exactly that faith. Anthropic itself has leaned on measures meant to be checkable, from watermarking what its models write to publishing its safety thresholds, though critics would say a company grading its own homework is part of the trust problem rather than the cure.

The wider industry is circling the same worry. Mark Zuckerberg has argued against a few firms owning the most powerful AI, and financiers have started organising around the risks of getting it wrong. Different framings, same underlying anxiety: a lot of power is accruing to a small number of companies, quickly, and most people did not get a vote.

Amodei is probably right that this is about trust. The harder question is whether an AI lab, of all institutions, is well placed to rebuild it.