Anthropic CEO Dario Amodei has a surprisingly simple prescription for fixing the public's growing distrust of AI:
Actually deliver something extraordinary.
In a series of posts this weekend, Amodei pushed back against criticism that his warnings about AI's dangers have helped create the backlash now facing the industry.
His argument is that the problem runs deeper than messaging.
People don't trust AI companies.
And, according to Amodei, the only convincing way to change that is to demonstrate that the technology can deliver meaningful benefits.
His example was deliberately blunt: actually curing cancer.
Stop Promising. Start Delivering.
Amodei acknowledged that saying AI will cure cancer has become almost meaningless.
"It's a cliché," he argued.
The problem, he says, is that AI companies have made enormous promises about how the technology will transform the world without yet delivering enough tangible results to justify that confidence.
That makes the criticism of the industry fair, in his view.
The solution isn't necessarily better PR.
It's results.
Anthropic is therefore increasing its work in biological and medical applications, with Amodei saying he hopes the company will have significant results in the coming years and early indications in the near term.
If those breakthroughs arrive, he believes the public will notice.
Is AI's Reputation Really About Messaging?
Amodei's comments came after investor Gavin Baker argued that his warnings about AI risk have contributed to growing public opposition to the technology, including resistance to data-center construction.
Amodei rejects that explanation.
He says his public writing has been roughly balanced between AI's risks and its potential benefits, pointing to his 2024 essay Machines of Loving Grace as an example of his efforts to describe an optimistic future for AI.
Instead, he believes the underlying problem is much broader.
It's a crisis of trust.
People are skeptical of corporations, governments and the technology industry generally, Amodei argues, and AI has simply become the latest place where that distrust is being expressed.
That interpretation is important because it suggests the industry's communications problem can't necessarily be solved by hiring better PR people or changing the tone of AI CEOs.
But Breakthroughs May Not Be Enough
There is an obvious challenge to Amodei's argument.
Even industries that deliver extraordinary benefits aren't automatically trusted.
Angel Brodin, an AI architect at OpenAI, pointed out that the pharmaceutical industry routinely produces major advances in public health while remaining one of America's least trusted industries.
Her argument is that people judge technology companies on more than what they invent.
They also care about pricing, access, lobbying, transparency, distribution of economic gains and who ultimately holds the power.
That may be the more difficult problem for AI.
A cancer cure could demonstrate that AI is capable of extraordinary things.
It wouldn't necessarily answer the question of who controls the technology, who benefits from it and who pays for it.
The Open-Source Problem
That debate also explains some of the criticism directed at Anthropic itself.
The company has chosen not to open-source its frontier models, citing safety concerns.
Former Meta chief AI scientist Yann LeCun argues that this kind of closed development contributes to public distrust and says AI should instead become widely available and shared.
Amodei disagrees with the idea that open weights alone solve the concentration problem.
His argument is that AI is structurally inclined toward concentrating power because the most advanced systems require enormous amounts of computing power and access to chips.
In his view, carefully designed regulation could potentially reduce risks while also limiting the power of the largest AI companies and leaving room for open-weight models.
That is a much more complicated proposition than simply open AI versus closed AI.
What This Means for Miami
Miami is becoming an increasingly important market for AI infrastructure, investment and adoption, which makes the question of public trust more than an abstract debate.
The city is also experiencing the physical consequences of AI expansion, from data-center demand to growing interest in AI-powered businesses and services.
If Amodei is right, the industry's credibility won't ultimately be determined by how convincingly companies explain AI.
It will be determined by what AI actually delivers.
And that leaves the industry with a rather uncomfortable challenge:
If you want people to trust AI, eventually you have to give them a reason to.
Reporting Source: This article builds upon reporting from Business Insider and TechCrunch and adds analysis of what the development means for Miami and South Florida.