Anthropic CEO Dario Amodei rejected the assertion that his warnings about artificial intelligence are the main cause of public backlash and regulatory pressure. The critique came from investor Gavin Baker, who said on the All-In podcast and on X that Amodei’s comments had helped stir opposition to AI, especially around data centers, and urged a more positive industry tone.
Amodei’s response
Amodei wrote that his public messaging has been “about equally balanced between risks and benefits,” noting he authored one major essay on benefits and another on risks. He said he penned the essay “Machines of Loving Grace” because he felt the industry was not articulating a sufficiently inspiring vision of AI’s potential.
At the same time, Amodei acknowledged a widespread negative view of AI and described it as “fundamentally a crisis of trust.” He argued that ordinary people distrust companies, governments and the tech industry and suspect those actors of seeking ways to harm them.
Amodei also accepted a degree of responsibility, saying the most accurate criticism of AI companies is that they have not yet delivered on big promises to benefit the world. He characterized claims that AI will cure cancer as “more a cliche than it is inspiring,” and said actually curing cancer would be the sort of achievement that could change public opinion.
Regulation, concentration and open-weights
Responding to Baker’s framing of regulation as a choice between broad distribution and concentration of power in a few firms, Amodei called that a “false choice.” He argued that the shorthand equating regulation with regulatory capture and concentration is overly simple and that many people outside Silicon Valley see regulation as a constraint on corporate power.
Amodei said Anthropic’s policy proposals are crafted to slow frontier AI companies while advantaging smaller competitors. He described AI as a technology that “tends to concentrate power” and said open-weights help but are not sufficient, because they shift advantage toward those with the most compute and chips.
Amodei argued the right rules could address cyber, bio and alignment risks, limit the institutional power of frontier firms, and still leave room for open-weights models while managing their specific risks. His posts were updated to include additional quotes, including material about open-weights, and a related tweet was posted on August 15, 2026.
Original source: TechCrunch AI