How the exchange began

Anthropic CEO Dario Amodei said the growing backlash against AI is "fundamentally a crisis of trust". The comments came after criticism directed at him by the investor Gavin Baker.

Baker's claim was that Amodei's warnings about the risks of AI had strengthened the backlash in the US against technology, and against data centers in particular. Baker also argued that Amodei had lost the debate on AI regulation and that, as the head of one of the sector's leading companies, he should adopt a more positive approach.

Amodei's answer

In statements published on X, Amodei disagreed with the view that his messaging is excessively negative. He said he takes a balanced approach between the risks and the potential benefits.

As grounds for that defence he pointed to his own essay, "Machines of Loving Grace". According to Amodei, he wrote that piece precisely because he felt no sufficiently inspiring picture had been drawn of how AI could transform the world for the better.

The regulatory side

The concrete ground of the argument is regulation. The regulations Anthropic supports include California bills that would impose transparency obligations on large AI companies.

The two positions are worth separating:

  • Baker's case: risk rhetoric hardened public opinion, and an industry leader should speak more positively.
  • Amodei's case: the source of the backlash is not the rhetoric but existing distrust of technology.
  • What stands out: the party accusing Amodei of going too far on regulation is not a regulator but an investor.

The debate, in other words, is not running on whether rules should exist, but on communication choices within the industry itself.

Why it matters

Amodei's thesis draws a distinction about the source of the backlash: people are wary of AI not because they heard industry leaders' risk warnings, but because they already distrust the technology sector. If that is right, stopping the warnings would not reduce the backlash; it would only leave the industry unable to discuss its own risks.

The opposing case is also arguable: an AI company CEO describing the dangers of his own product strengthens the public question of "if it is dangerous, why are you building it", and that question travels down to the local councils where data center permits are debated.

There is no measurable side to this argument; it is a choice between two narratives that both look reasonable. What is notable is that the debate is no longer running between regulators and the industry, but inside the industry. That the investor side wants to manage public perception like an asset also shows how dependent AI company valuations have become on that perception.