Anthropic CEO Dario Amodei has pushed back against criticism that warnings about AI risks are fuelling public negativity. In a post shared on social media platform X (formerly known as Twitter) Amodei said that the real issue is a âcrisis of trustâ rooted in decades of suspicion toward companies, governments, and the tech industry. âOrdinary people donât trust us and suspect weâre cooking up some new way to screw them over,â he wrote.

Anthropic CEO Dario Amodei dismissed calls for AI firms to rely on glossy campaigns

Amodei dismissed calls for AI firms to rely on glossy campaigns to improve their image. âAt this point, saying that AI will cure cancer is more a cliché than inspiring,â he said. Instead, he argued that the only way to rebuild trust is to actually deliver breakthroughs like curing cancer, rather than making promises that feel deceptive.

Highlighting his own approach, Amodei said he is accountable to biomarkers and adjusts protocols based on measurable outcomes. He compared mastery in diet and health to excellence in academics, sports, or engineering, insisting that selfâcontrol and discipline should be respected as legitimate forms of mastery.

âI do agree that the public has a negative view of AI (and that this is a big problem), but I donât think it is primarily caused by me or any other AI leader warning about AIâs risks.

I think it is fundamentally a crisis of trust. I think that ordinary people donât trust companies, governments, or the tech industry and always suspect that we are cooking up some new way to screw them over. The causes of this go back decades and AI is just the latest iteration of it. I donât think that a glitzy marketing campaign with a positive spin (which some have advocated that Anthropic do) is the way to win back that trust â at this point, saying that AI will cure cancer is more a cliche than it is inspiring, and most people think it is deceptive. The thing that will work is actually curing cancer. I think by far the most accurate criticism of AI companies including Anthropic is that we havenât yet delivered on our big promises to benefit the world. That is totally on us, and I think itâs the criticism you should be making, instead of all this stuff about messaging and marketingâ, wrote Amodei on X.

A rare, lengthy social media post

Amodei laid out his argument in an extended post on X on Saturday, a platform he generally avoids, according to a Fortune report. He began by addressing the debate around AI regulation, describing what he sees as a false choice between critics who worry regulation leads to concentrated corporate power and those who argue that widely distributing AI, including through open-weight models, is the better safeguard. Amodei argued that institutions like the court system show power can be decentralized without abandoning oversight, and noted that Anthropic has backed policies that would slow down frontier AI companies like itself while giving smaller competitors more room to operate.

He acknowledged that AI does tend to concentrate power structurally, but attributed that tendency to the scaling laws underlying how AI models improve as more computing resources are put into them, rather than to regulation itself. In his view, open-weight models only shift that concentration toward whoever controls the most computing capacity and chips, rather than eliminating it. He said he favors "rules of the road" that address AI's cyber, biological and alignment risks while also institutionally constraining the power of frontier AI companies and leaving room for open-weight models, and voiced support for a FINRA-like regulatory body as well as the Trump administration's approach to AI testing.

What comes next, according to Dario Amodei

Amodei said Anthropic is trying to rebuild credibility by producing tangible results, pointing specifically to increased work in biology and medicine, and describing "early glimmers" of what he suggested could eventually be significant breakthroughs. Until there are actual accomplishments to point to, he said the company would avoid making further big promises and instead focus on being transparent about AI's risks in the meantime, arguing that honesty carries no more public credibility risk, and quite possibly less, than an approach that downplays or distracts from risks people already sense are real.