Anthropic: Amodei Links AI Rejection to a Crisis of Trust

Dario Amodei contests the idea that he paints a bleak picture of AI and attributes the current backlash to a deeper distrust of companies and public authorities. He advocates for rules that could address security risks, regulate leading companies, and preserve a space for open-weight models. Investor Gavin Baker, on the other hand, accuses him of fueling rejection and losing the battle of ideas on regulation, a claim that the leader refutes.

Amodei defends rules that hinder leading companies

Dario Amodei believes that presenting the choice between unregulated AI and a technology concentrated by regulation is a false dilemma. He describes a "jargon" from Silicon Valley that equates regulation with regulatory capture and concentration of power, a view he considers overly simplistic, noting that many actors outside this bubble see regulation as a means to constrain corporate power and benefit ordinary people. While he does not fully embrace this latter position, he asserts that Anthropic carefully formulates its policy proposals.

He aims to design rules that disadvantage—thus slow down—leading AI companies while benefiting smaller rivals. According to him, AI structurally tends to concentrate power; open-weight models would only partially change this by shifting concentration toward those who possess the most computing power and chips. He argues that good "rules of the road" can simultaneously address cyber, bio, and alignment risks, institutionally limit the power of leading companies, and leave room for open-weight models by managing their specific risks.

The rejection of AI stems from a crisis of trust, according to Amodei

Dario Amodei characterizes the backlash against AI as a crisis of trust. He describes a public that does not trust companies, governments, or the tech industry, and suspects that they are being deceived. This distrust, he claims, has been rooted for decades, with AI being merely the latest manifestation.

He acknowledges that the perception of AI is negative and considers it a major problem, while contesting that this attitude primarily stems from his own warnings or those of other industry leaders. Trust has indeed become a recurring theme in the debate surrounding AI CEOs.

Unfulfilled promises and a striking example

For Dario Amodei, the most pertinent criticism directed at AI companies, including Anthropic, lies in the gap between grand promises and the actual benefits delivered to the world. He takes responsibility for this and argues that the issue of real outcomes should take precedence over discussions of messaging and marketing.

He cites a telling example: promising that AI will cure cancer is a cliché, and it is the actual curing of cancer that, according to him, would shift public opinion.

Baker accuses Amodei of negativism, the leader responds

Investor Gavin Baker argues that Dario Amodei's warnings have fueled a rejection of AI in the United States, particularly against data centers. He elaborated on this thesis on the All-In podcast and on X, also claiming that Amodei has lost the argument on regulation and urging him to be a more positive advocate for his industry, while presenting him as poised to lead one of the most important companies in the world.

The debate occurs in a context where Anthropic has advocated for rules, including a California proposal imposing transparency obligations on major AI players. As CEO of Anthropic, Dario Amodei recently responded to the accusation of negativism: he rejects the idea that his message is disproportionately dark, claims to write in a balanced manner between risks and benefits, and recalls having published "Machines of Loving Grace" to offer a more inspiring vision of AI's potential contributions.