Anthropic CEO Says AI Has a Trust Problem And Tech Companies Have Themselves to Blame

Anthropic CEO Dario Amodei says growing public resistance to artificial intelligence cannot simply be blamed on warnings about the technology's dangers.

The deeper problem, he argues, is trust.

Amodei made the case while responding to investor Gavin Baker, who accused him of contributing to America's growing AI backlash by repeatedly emphasising risks associated with increasingly powerful systems.

Baker argued that warnings from influential AI executives have helped strengthen opposition to everything from AI regulation to the enormous data centres required to power the technology.

He also suggested that Amodei should present a more optimistic picture of an industry in which his company is now one of the world's most influential players.

Amodei rejected that characterization.

The Anthropic chief argued that his public writing has consistently addressed both the potential benefits and dangers of advanced AI.

One example is his essay Machines of Loving Grace, in which he explored how sufficiently capable AI could accelerate progress in areas including biology, medicine, neuroscience, economic development and governance.

But Amodei accepts one important part of the criticism: a significant section of the public does not like where AI appears to be heading.

He just disagrees about why.

“I think there is a crisis of trust,” Amodei said, arguing that many people have become deeply suspicious of companies, governments and technology institutions and increasingly assume that powerful organisations are acting in their own interests.

For the AI industry, that creates a problem that better marketing may not solve.

Amodei argued that technology companies ultimately need to demonstrate that their systems can deliver meaningful improvements to people's lives rather than continually making enormous promises about what AI might eventually accomplish.

He used medicine to illustrate the point.

Talking about AI eventually helping to cure diseases is easy. Actually contributing to breakthroughs that improve or save people's lives would do considerably more to change public attitudes.

That distinction matters as AI companies pour billions of dollars into increasingly powerful models and the infrastructure needed to operate them.

The rapid construction of data centres has itself become politically contentious in parts of the United States, where communities have raised questions about electricity consumption, water use, environmental effects and whether the economic benefits justify the infrastructure demands.

Amodei's argument also extends to regulation.

Anthropic has supported rules requiring greater transparency and safeguards from developers of the most powerful AI systems. The company has previously set out its position on regulating frontier artificial-intelligence models, arguing for requirements that increase as systems become more capable.

Critics of that approach contend that expensive compliance requirements could ultimately strengthen the largest AI companies because smaller competitors may struggle to meet them.

Amodei rejects the idea that regulation must automatically produce that result.

He argues that carefully designed rules can deliberately impose greater obligations on frontier AI companies while leaving more room for smaller developers and open-weight models.

At the same time, he believes AI has characteristics that naturally encourage concentrations of power.

Building the world's most advanced models requires enormous quantities of computing power, specialised chips, data, engineering expertise and capital. Even making model weights openly available does not entirely solve the problem if only a limited number of organisations possess the computing infrastructure required to exploit them at scale.

For Amodei, the challenge is therefore not choosing between unrestricted AI development and regulation.

It is designing rules capable of addressing serious risks — including cybersecurity, biological misuse and loss of control over advanced systems — without allowing a handful of companies to dominate the technology.

His broader argument leaves the AI industry with an uncomfortable challenge.

The companies building increasingly powerful systems can tell people that AI will transform medicine, education, productivity and scientific discovery.

But trust will ultimately depend less on what the industry promises than on what ordinary people actually experience.

And on that measure, Amodei's message to his own industry is straightforward: it still has something to prove.