Your Own Experts Are Frightened: What the AI Insider Warnings Reveal About Your Dissent Deficit
When the people closest to the capability sound the alarm and the people governing it shrug, you don't have a technology problem. You have a governance one.
Anthropic's chief executive wants AI development slowed. His former researchers say staff are genuinely frightened. Silicon Valley's response has been a collective shrug, and the White House has denounced the whole enquiry as hysteria. Strip away the subject matter and you are left with the oldest governance failure in commerce: expert warning, executive dismissal, and no mechanism connecting the two.
A former Anthropic researcher told the BBC this week that staff inside the company are "genuinely frightened" for humanity's future. The firm's own chief executive, Dario Amodei, has called for the pace of AI development to be slowed on the grounds that the risks are serious. And the response from much of Silicon Valley — and from the President of the United States, who denounced calls for regulation outright — has been somewhere between scepticism and contempt.
Here is what should interest you, and it has almost nothing to do with artificial intelligence.
The people closest to the capability are raising alarms. The people with the authority to act are discounting them. And there is no institutional mechanism that reliably converts the first group's judgement into the second group's decisions. That is not an AI story. That is a description of how most large organisations handle inconvenient expertise, and it is playing out inside your business right now on a smaller stage with less press coverage.