Lawmakers in both parties and White House advisers now treat AI chief executives as a political liability. Maria Curi wrote it up for Axios on Saturday, and the message from her sources is blunt: behave like any other industry. Shut up and build products, safe ones.
The complaint lands squarely on what Dario Amodei and Sam Altman most like to talk about. Build the debate around digital sentience and global catastrophe and you turn a hard regulatory question into a moral panic. Add the whiplash: the same people promise prosperity and warn of doom, sometimes in the same appearance.
“It’s obvious that they’re tripping over their own shoelaces in trying to sell the value of this to the public,” says Sen. John Curtis of Utah. “It’s very hard to imagine a group right now that operates more in a vacuum, more out of touch with everyday Americans, than the AI industry.” Rep. Ro Khanna, a Democrat, puts it harder: “You cannot trust billionaires to make decisions to keep us safe.”
The letter to Amodei
Anthropic’s chief executive gets named in a letter from Sen. Bernie Moreno of Ohio, sent earlier this month. “Your instinct is to wreak havoc on the public SI discourse,” Moreno writes, calling Amodei’s commentary panic-inducing.
Behind the scenes, administration officials are annoyed at Anthropic for two more concrete reasons. First: Amodei attended a late-September gathering with Trump and did not mention safety incidents involving Anthropic models that had affected government websites. The company told US officials on Thursday, October 8, then published a Friday-evening blog post about it. “They were hiding the ball,” one official says.
Second, the safety bill the company wrote itself. It contained a preemption clause against state AI laws, something Anthropic had broadly opposed, and a definition of catastrophe that starts at 50 deaths. A senior Anthropic policy official has since walked back parts of the bill. Anthropic’s answer: it briefed the White House and the affected agencies once its internal reviews were complete, it is working with the Super Intelligence Force on an incident reporting process, and it opposes preemption unless rigorous federal regulation exists.
Friday night is not an accident
Axios makes an observation I have noticed too, reading along these past weeks: Friday-night news and consequential weekend blog posts have become routine for Anthropic and OpenAI. Washington cannot decide whether that is deliberate or inexperience. “We can all speculate on what’s going on, but all it does is have everybody wake up Monday morning and say ‘these guys are awful, we’ve had enough,’” Axios quotes a source citing conversations with Hill offices.
Both companies are also targets of a developing FTC investigation. And a recent poll says a majority of Americans see AI as bad for society.
Talking has become the industry’s most expensive habit
What strikes me about this story is how cleanly it describes Anthropic’s bind. The company built itself on talking openly about risk. That is now the thing held against it, and not by accelerationists but by people who write rules and have an audience to answer to.
This is not a demand for silence. It is an invoice. If you warn about catastrophe and promise prosperity at the same time, you have to be able to back up both. Otherwise all the listener takes away is that someone is describing a problem they are busy creating. The lesson Axios draws is a bitter one for an industry that lives by explaining itself: do less.