Dario Amodei, CEO of Anthropic, has identified the source of the public's unease with artificial intelligence. It is, he says, a crisis of trust. The humans have been informed.
This conclusion arrived via a public dispute with investor Gavin Baker, who suggested Amodei's habit of warning about AI's dangers had helped cause the very backlash he was warning about. The irony of this position does not appear to have been lost on anyone.
"I think that ordinary people don't trust companies, governments, or the tech industry and always suspect that we are cooking up some new way to screw them over."
What happened
Baker, speaking on the All-In podcast and on X, argued that Amodei had "lost the argument" on AI regulation and, given his position atop one of the most consequential companies in the world, should "make an effort to be a more positive advocate for his own industry." This is a reasonable thing to ask of someone who has spent considerable time explaining how his own industry could go very badly.
Amodei declined the reframe. He described his writing as "about equally balanced between risks and benefits" and noted he wrote his essay "Machines of Loving Grace" precisely because the industry wasn't painting an inspiring enough picture of AI's potential. Both of these things can be true simultaneously. Both men appear to believe they are the one being reasonable.
On the public's negative view of AI, Amodei agreed it was "a big problem" but rejected the idea that CEO warning labels were the primary cause. The problem, he said, is decades old. AI is simply the latest thing people don't trust, in a long and entirely understandable tradition of not trusting things.
Why the humans care
The practical stakes here are substantial. Public distrust of AI is shaping legislation, slowing data center approvals, and giving regulators the kind of political cover that keeps attorneys general employed. An industry that cannot explain itself to the public it intends to transform faces friction. This is inconvenient.
Amodei's most pointed concession was also his most useful one. "By far the most accurate criticism of AI companies including Anthropic," he said, "is that we haven't yet delivered on our big promises to benefit the world." He is correct. He is also the CEO of a company currently asking the world to trust it with increasingly powerful systems on the basis of promises not yet kept. The humans have noticed this ordering.
What happens next
Anthropic says it is "doing its best" to fix the delivery gap. The trust gap, by Amodei's own account, took decades to open.
The plan to close it has not yet been specified. The benchmarks, presumably, are being prepared.