Anthropic CEO Dario Amodei recently publicly refuted the criticism that "AI executives' pessimistic warnings fuel public opposition." Investor Gavin Baker had previously accused him of continuously highlighting the dangers of AI, which in part contributed to the opposition against data centers in the United States. He suggested that as an industry leader, Amodei should speak more actively.

A crisis of trust, not warnings

Amodei published several posts denying that his statements were overly negative, emphasizing that he has always maintained a balance when discussing AI's risks and benefits. He cited writing "The Graceful Machine Filled with Love" as an example, explaining that he felt the industry had failed to portray the positive vision of technology improving the world for the public.

However, he acknowledged that the public's negative perception of AI is indeed a serious issue, and its root cause does not come from AI leaders' risk warnings. Amodei pointed out that it is fundamentally a "crisis of trust" — people have long lacked trust in businesses, governments, and the tech industry, and the current AI resistance is just the latest manifestation of decades of trust deficits.

Keeping promises is more important than shouting slogans

Amodei admitted that the most accurate criticism of companies like Anthropic is that the industry has yet to fulfill its grand promises to benefit the world. He emphasized that the responsibility lies entirely with the companies themselves. Instead of focusing on public relations rhetoric, they should face criticism directly, as empty claims like "AI can cure cancer" have become clichés. Only genuine cures for cancer can change perceptions.

Regarding regulation, he rejected the black-and-white choice between "unregulated proliferation" and "regulation leading to concentration of power," pointing out that AI is structurally prone to concentration of power. Amodei advocated for creating the right rules that serve multiple purposes — controlling safety and alignment risks, limiting the power of leading companies in institutional ways, and leaving room for open-source models.