Anthropic's CEO admitted: The AI industry has failed to fulfill its promises, and this is the root of the trust crisis.
The key complaint against the artificial intelligence industry is not fears of singularity or a machine uprising, but the simple failure to deliver on loud promises. This statement was made by Anthropic CEO Dario Amodei, linking society's cooling toward the technology to a deep, systemic crisis of trust that has been building for decades.
The occasion for the discussion was a post by investor Gavin Baker, who accused Amodei of overly grim rhetoric. In Baker's view, the head of a startup preparing for status as one of the world's most important public companies should adopt a more positive tone. Amodei responded with a lengthy series of posts on X, emphasizing that marketing tricks can no longer solve the problem.
The core of the dispute: what matters more—words or deeds?
Amodei insists that he pays equal attention to both the risks and the benefits of the technology. As an example of a positive agenda, he recalled his essay Machines of Loving Grace, which describes the potential of powerful AI for medicine and the economy. However, the main point of his response is an admission of the industry's own guilt.
"I think at the core there is a fundamental crisis of trust. Ordinary people do not trust companies, governments, or the technology industry and always suspect that we are coming up with some new way to deceive them."
According to him, the most fair criticism of Anthropic and its peers is not talk about safety, but the absence of real results. "This is entirely on us. This is what we should be criticized for, not for messaging and marketing," Amodei wrote.
"Only one thing will work—actually curing cancer"
The head of Anthropic believes that promises about an AI revolution in healthcare have become a cliché and are perceived as deception. The situation can only be corrected through actions. He announced intensified work in biology and medicine, promising "incredible results" in the coming years, although he did not provide details.
However, not everyone agrees with this vision. Renowned researcher Yann LeCun responded by stating that for a decade he has been promoting the only correct path—open and accessible technology. By his analogy, AI, like the printing press or the internet, amplifies human intelligence, and therefore we need a diversity of systems, not their concentration. He also cited Mark Zuckerberg's position, who considers problematic the very idea of AI being so dangerous that extreme centralization of power is required.
This approach directly contradicts Anthropic's policy, which, while not advocating a complete ban on open models, insists on stricter safety requirements as system capabilities grow.
A separate viewpoint was expressed by OpenAI GTM specialist Angel Brodin. She recalled the pharmaceutical industry: real breakthroughs in medicine do not save it from low levels of trust. In her view, attitudes toward AI will depend not only on technology but also on pricing, transparency, and the distribution of economic benefits.
My analysis: Amodei is certainly right that the industry has overheated expectations. But his thesis that "just curing cancer is enough" oversimplifies the problem. Trust is a complex product that includes not only scientific victories but also implementation ethics, social responsibility, and a willingness to share power. While AI giants argue about risks, society is waiting not for words but for clear rules of the game that will make the technology safe and useful for everyone, not just shareholders.