Anthropic CEO Dario Amodei has pushed back against claims that his warnings about artificial intelligence have contributed to growing public opposition to the technology, explaining that the industry’s deeper problem is a longstanding “crisis of trust” rather than the way AI executives communicate its risks.
Amodei was responding to investor Gavin Baker, who said on the All-In podcast and on X that the Anthropic chief’s warnings about AI’s potential dangers have helped fuel resistance to the technology in the United States, including opposition to the rapid expansion of data centers.
Baker said Amodei had “lost the argument” over AI regulation and argued that, as he prepares to lead what could become one of the world’s most important companies, he should adopt a more positive public stance toward the industry.
“I respectfully think he should make an effort to be a more positive advocate for his own industry,” Baker wrote.
Amodei rejected the premise that his public messaging has been disproportionately negative. He said his writing has been “about equally balanced between risks and benefits” and pointed to his essay “Machines of Loving Grace” as evidence that he has also tried to articulate an optimistic vision for AI.
Amodei said he wrote the essay because he believed the AI industry was not presenting “an inspiring enough picture of how the technology could radically transform the world for the better.”
He nevertheless acknowledged that the public’s perception of AI remains negative.
“The public has a negative view of AI,” Amodei said, adding that “this is a big problem.”
But he disputed the argument that public skepticism is primarily the result of warnings from him or other AI executives.
“I think it is fundamentally a crisis of trust,” Amodei said. “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.”
His argument places the backlash against AI within a broader deterioration of public confidence in large institutions rather than treating it as a communications problem created by the technology industry itself.
Amodei said that distrust had developed over decades and that the current backlash against AI is “just the latest iteration of it.”
AI Companies Must Deliver, Not Promise
For Amodei, the strongest criticism of the AI industry is not that executives have been too pessimistic but that companies have yet to demonstrate enough tangible benefits to justify their promises.
“I think by far the most accurate criticism of AI companies including Anthropic is that we haven’t yet delivered on our big promises to benefit the world,” he said.
“That is totally on us,” Amodei added, arguing that this is the criticism the industry should face rather than scrutiny of its messaging and marketing.
He distinguished between promising transformative outcomes and actually delivering them. Saying that AI could cure cancer, for example, does little to persuade a skeptical public if those benefits remain theoretical.
Promising that AI will cure cancer is “more a cliche than it is inspiring,” Amodei said. What would change people’s views, he argued, would be “actually curing cancer.”
The comments highlight a central challenge for AI companies as they attempt to maintain public support while making increasingly ambitious claims about the technology’s economic and scientific potential.
The industry has promoted AI as a tool that could accelerate scientific discovery, improve productivity and transform healthcare, while simultaneously warning about risks involving cybersecurity, biological threats, misinformation and autonomous AI systems. That combination has made the industry’s public messaging unusually difficult. Companies must convince investors and customers that AI will produce enormous benefits while persuading governments and the public that powerful systems can be deployed safely.
Amodei Rejects Regulation-Versus-Innovation Argument
Amodei also challenged Baker’s characterization of the debate over AI regulation. Baker had argued that tighter regulation could concentrate AI development in the hands of a small number of dominant companies by making it more difficult for smaller competitors to operate.
Amodei called that a “false choice” between distributing AI without regulation and concentrating the technology through regulation.
He acknowledged that Silicon Valley often treats regulation as synonymous with regulatory capture and greater concentration of power, but said that view is too simplistic.
“I know that there’s a sort of Silicon Valley shorthand where regulation = regulatory capture = concentration of power, but I’ve always found this to be an overly simplified picture of the world,” Amodei said.
“Many people outside this bubble think of regulation as something that constrains corporate power and benefits ordinary people,” he added.
Amodei said he does not necessarily share that view in every circumstance, but argued that it explains why Anthropic has approached its policy proposals cautiously.
The company’s position is not simply that frontier AI companies should face more regulation. Amodei said Anthropic tries to develop proposals that would slow or disadvantage the largest AI companies while creating room for smaller competitors.
“We try very hard to make proposals that disadvantage (slow down) frontier AI companies while advantaging smaller competitors,” he said.
Anthropic has supported some AI regulations, including a California bill requiring transparency from large AI companies.
The argument goes to a broader question about the structure of the AI industry.
Amodei said AI is “structurally” a technology that tends to concentrate power because developing the most capable systems requires access to enormous amounts of computing capacity, advanced chips, capital and specialized talent. Open-weight models can distribute some capabilities more broadly, he said, but they do not eliminate the underlying concentration.
“Open-weights do help some with this but are nowhere near a sufficient solution because they simply shift the concentration somewhat to those with the most compute and chip,” Amodei said.
That argument challenges the idea that simply making powerful models openly available would solve concerns about concentration in AI.
Companies with access to the largest computing clusters and most advanced semiconductor technology would still possess significant advantages in developing and deploying increasingly capable systems.
Amodei said instead for a regulatory framework that establishes what he called the “rules of the road” for the industry.
He said the right framework should simultaneously address AI’s cybersecurity, biological and alignment risks, constrain the institutional power of frontier AI companies and preserve room for open-weight models while addressing the specific risks associated with them.
Nevertheless, Amodei’s comments illustrate the difficult position Anthropic occupies in the AI market. The company is competing directly with OpenAI, Google and other major AI developers while simultaneously advocating for policies that could impose additional constraints on frontier AI development.
That position has generated criticism from parts of the technology industry, where some executives and investors believe that regulation could entrench today’s largest AI companies by raising compliance costs and making it harder for new entrants to compete. Amodei’s response is that carefully designed rules can do the opposite if they impose greater constraints on the largest companies while reducing barriers for smaller competitors.
The debate is likely to intensify as AI systems become more capable and their economic footprint expands. The industry is simultaneously seeking enormous investment in data centers and computing infrastructure, lobbying governments over AI policy and attempting to persuade the public that increasingly powerful systems can be trusted.
For Amodei, the solution to public skepticism is not simply more optimistic messaging.
It is performance.
The industry’s credibility, he notes, will ultimately depend on whether AI companies can deliver measurable benefits that ordinary people can see and use, while demonstrating that the risks created by powerful systems can be managed.
