Anthropic CEO Dario Amodei says the AI industry has a trust problem — and the best way to fix it is by delivering real-world breakthroughs instead of making bigger promises.
In a rare post on X, Amodei acknowledged growing public skepticism toward artificial intelligence and admitted that AI companies, including Anthropic, have yet to fully deliver on their bold claims about how the technology could benefit society.
He argued that simply saying AI could one day cure diseases like cancer is no longer convincing. Instead, companies need to produce tangible results that demonstrate the technology’s value.
Amodei’s comments come as the AI industry faces mounting criticism over massive data centers, the use of copyrighted material for model training, potential job losses, and broader concerns about how rapidly the technology is developing.
Amodei himself has frequently warned about AI risks. He has previously raised concerns about cybersecurity, financial systems, critical infrastructure, national security, and the possibility that AI could eliminate large numbers of entry-level jobs.
However, he rejected the idea that warnings from AI leaders are the main reason the public distrusts the industry. Instead, he suggested that skepticism toward AI is part of a broader decline in trust toward large corporations.
Anthropic has also faced criticism for keeping its most advanced models closed-source. Former Meta AI chief scientist Yann LeCun responded to Amodei by arguing that making AI more open, widely available, and diverse would do more to build trust.
Others questioned whether scientific breakthroughs alone would be enough. OpenAI applied AI architect Angel Brodin pointed to the pharmaceutical industry, arguing that even major health advancements do not automatically create public trust. She said people will also judge AI companies on factors such as pricing, access, transparency, lobbying, economic inequality, and who ultimately controls the technology.
Amodei said Anthropic is increasing its work in biology and medicine and expects early signs of progress in the coming months, with potentially much larger results in the years ahead.
His message was simple: if AI companies want the public to believe their promises, they eventually need to prove them.