Anthropic CEO Says AI Backlash Is a Crisis of Trust, Not His Fault
So Dario Amodei, the CEO of Anthropic, is pushing back against claims that he’s been too negative about AI. And honestly? He’s got some interesting points.
Let me break down what’s happening here.
What’s the Drama?
An investor named Gavin Baker went after Amodei recently. Baker argued on the All-In podcast and on X that Amodei’s warnings about AI dangers have helped fuel a backlash against AI, especially against data centers.
Baker basically said: “You’ve lost the argument on AI regulation. You’re about to be the CEO of one of the most important companies in the world. Maybe try being a more positive advocate for your own industry?”
Pretty direct, right?
What Amodei Said Back
Amodei responded with a series of posts. And here’s what he said.
First, he pushed back on the idea that his messaging has been disproportionately negative. He says his writing has been equally balanced between risks and benefits. He even pointed to his essay “Machines of Loving Grace” where he tried to paint an inspiring picture of how AI could radically transform the world for the better.
But here’s where it gets interesting.
Amodei agreed that the public has a negative view of AI. And he agreed that this is a big problem. Where he disagreed was with the idea that he or other AI leaders are primarily causing that negativity.
He called it “fundamentally a crisis of trust.”
The Trust Problem
Here’s what Amodei said about trust:
“I think ordinary people don’t trust companies, governments, or the tech industry and always suspect that we are cooking up some new way to screw them over.”
He says this crisis of trust has been decades in the making. The AI backlash is just the latest version of it.
And honestly? He might have a point. People have been burned by tech companies before. Privacy issues. Data breaches. Surveillance. It’s not hard to see why people are skeptical.
You might be thinking: “Didn’t Anthropic just do something with watermarks?”
Yep. Literally last week. They announced invisible text watermarks on Claude’s output to comply with EU regulations. I wrote about what it actually means, whether it affects quality, if code gets watermarked, and whether people can remove it. Check it out here.
The Real Criticism
Amodei said something that really stood out to me.
“I think by far the most accurate criticism of AI companies including Anthropic is that we haven’t yet delivered on our big promises to benefit the world. That is totally on us.”
His point? Stop talking about the messaging and marketing. The real criticism is that AI companies keep promising to cure cancer and save the world, but haven’t actually done it yet.
He put it this way. Promising AI will cure cancer is more of a cliche than it is inspiring. What would actually change people’s opinions? Actually curing cancer.
Fair point.
What About Regulation?
This is where the debate gets more interesting.
Baker argued that Amodei’s support for regulation is basically creating a false choice. Either you distribute AI widely without regulation, or you concentrate AI power in a few companies through regulation.
Amodei disagreed with that framing.
He said: “I know that there’s a sort of Silicon Valley shorthand where regulation = regulatory capture = concentration of power, but I’ve always found this to be an overly simplified picture of the world. Many people outside this bubble think of regulation as something that constrains corporate power and benefits ordinary people.”
He doesn’t necessarily agree with that perspective either, but he says it’s why Anthropic has been very careful with their policy proposals.
His Actual Position on Regulation
Amodei said they try hard to make proposals that disadvantage frontier AI companies while advantaging smaller competitors.
Why?
He believes that “AI is structurally a technology that tends to concentrate power.” Even open-weights models don’t fully solve this problem because they just shift the concentration to whoever has the most compute and chips.
His ideal? Rules that can:
- Address AI risks around cybersecurity, bioweapons, and alignment
- Institutionally constrain the power of frontier AI companies
- Leave room for open-weights models while addressing their specific risks
That’s his position.
What This Really Means
Here’s the bottom line.
Amodei is saying: “The problem isn’t that I’m too negative. The problem is that people don’t trust tech companies. And they don’t trust us because we haven’t actually delivered on our promises yet.”
That’s a pretty self-aware take, honestly.
He’s not blaming the public for being skeptical. He’s saying the skepticism is earned. And the only way to fix it is to actually deliver results, not just talk about them.
Why This Matters
If you’re following AI, this is an important conversation.
The CEOs of major AI companies are caught between two forces. On one side, they need to warn about risks because AI really could be dangerous. On the other side, they need to be optimistic enough that people actually want to use their products.
Amodei is trying to walk that line. He’s admitting that the criticism is valid and the trust issues are real. But he’s also pushing back against the idea that he’s the reason people are skeptical.
The Bottom Line
Amodei says AI backlash is a crisis of trust, not a crisis caused by AI leaders warning about risks. He admits AI companies haven’t delivered on their promises yet. And he’s trying to find a path on regulation that doesn’t just concentrate power in the hands of a few companies.
He’s not wrong about any of that.
The question is whether people will start trusting AI companies when they actually start delivering. That remains to be seen.
One Comment