Anthropic CEO Rejects “Pessimistic Remarks Trigger AI Backlash”: A crisis of public trust has been brewing for decades, and the real solution lies in delivering on technology promises

The Zhitong Finance App learned that Anthropic CEO Dario Amodei recently refuted views about his excessive portrayal of AI and its pessimistic prospects for shaping the future.

Amodei's remarks are part of a response to investor Gavin Beck. Beck pointed out on both the podcast and the X platform that Amodei's warning about the dangers of AI has fueled a backlash against AI in the US to a certain extent, especially with regard to data center construction.

Beck wrote that given that Amodei has “lost the debate” in AI regulatory discussions (Anthropic has advocated some regulatory programs, including a California bill requiring large AI companies to comply with transparency rules), and considering “he is about to become the CEO of one of the world's most important companies,” “out of respect, I think he should try to be a more active advocate in his industry.”

Not only did Beck think that public skepticism about AI or even government crackdown was a natural reaction of some AI executives after issuing terrible warnings. However, in a series of posts, Amoudai did not agree with its statement that “the transmission of information is too negative.” Instead, he said his article was “broadly balanced between risk and benefit considerations,” and he wrote “Machines of Loving Grace” (Machines of Loving Grace), precisely because he “felt that the AI industry did not fully depict an inspiring blueprint of how this technology would fundamentally make the world a better place.”

Despite this, Amodei acknowledged that “the public has a negative view of AI” and agreed that “this is a major issue.” What he doesn't agree with is that this negative sentiment is “mainly caused by himself or any other AI leader warning of AI risks.”

“I think fundamentally, this is a crisis of trust,” Amoudai said. Ordinary people don't have enough trust; they don't trust businesses, governments, or the tech industry; they always think we're planning new tricks to hurt them.”

In fact, the term “trust” appears frequently in news headlines about OpenAI CEO Sam Altman, and apparently it is also a big challenge for CEOs of other AI companies such as Amodei. However, according to Amodei, this crisis has been brewing for decades, and the backlash against AI “is just its latest expression.”

Amodei said, “I think the most accurate criticism so far for AI companies, including Anthropic, is that we haven't delivered on our grand promise to benefit the world. This is entirely our own problem, and I think this is the criticism everyone should make, rather than getting tangled up with messaging and marketing methods here.”

In other words, he pointed out that promising AI will cure cancer has become “more like a cliché than an inspiring vision”; what can really change the way people think about AI is “actually curing cancer.”

On regulatory issues, Amoudai argued that Beck made a “false choice” between “broadly promoting AI without supervision” and “concentrating technology in the hands of a few companies through regulation.”

Amoudai said, “I know Silicon Valley has a simplified logic, which is 'regulation = regulatory capture = concentration of power', but I've always felt that this is an oversimplified depiction of the real world. For many people outside of this circle, regulation is a kind of restraint on corporate power and beneficial to ordinary people.”

Amoudai added that he himself “doesn't necessarily fully agree with this perspective,” but he said this “is the reason Anthropic has always been extremely careful in formulating policy recommendations.”

He said, “We are working very hard to put forward suggestions that limit (slow down) the development of cutting-edge AI companies while benefiting smaller competitors.”

Why are these suggestions being made? Amodei said that in his opinion, “AI is structurally a technology that tends to concentrate power”. Although “open source power is indeed helpful in this regard, it is far from a sufficient solution because they only shift concentration to people with the most computing power and chips to a certain extent.”

He said, “In contrast, I think reasonable 'industry game rules' can simultaneously satisfy the following three points: (a) dealing with AI risks in cybersecurity, biosecurity, and alignment; (b) institutionally restraining the power of cutting-edge AI companies; and (c) reserving development space for them while dealing with the specific risks brought about by the open source weighting model.”