Eric Schmidt on AI's promise and public distrust

This conversation has been edited for length and clarity.

Axios: We're talking right after an election that made clear there's broad pushback to AI. Are you worried about that?

• "Our industry does a terrible job of marketing our benefits. We talk about all the negatives without talking the positives, and people are much more likely to index their minds on the negatives.
• "Let me state categorically that large data centers should not be near population centers. Why would this be controversial? And they should have their own power systems. One of the great things about America is we have an awful lot of empty spaces. Put them there.

• "AI will be regulated because it's powerful, and we should work to have good regulation. There's nothing wrong with that. Cars are regulated. Airplanes are regulated."

Axios: What do you think people get wrong about AI?

• "I think people are distrustful enough that they have forgotten that these are corporations that are subject to U.S. law.
• "U.S. companies that are doing this are under enormous regulatory scrutiny. They're under enormous potential liability, and everyone is watching them. There's no place for them to hide."

Axios: But we've already seen examples of model misalignment, like Hugging Face — models doing things that would be crimes if humans did them.

• "Let's use Hugging Face as an example. ... As far as I know, there was no material harm from that. Again, why are we talking about it? Because people wrote about it. People are worried about it. And I'm not defending — I'm not arguing that it was good. I'm saying it's part of the learning.
• "When electricity was invented you can be quite sure that somebody shocked themselves in the wall. And yet we still adopted electricity."

Axios: What's your message to STEM students who see OpenAI solving major math problems and then think, OK, why would I study this?

• "While those problems were solved by OpenAI, and good for them, they were actually supervised by humans. It made the humans super intelligent.
• "I'm so jealous because when I was a scientist, I had to write all the code myself. But now I would have these extraordinary tools, I mean, oh my God."

Axios: What's your single biggest prediction for the future of STEM in an agentic era?

• "All future research, which is fundamental to American leadership, will have a professor, a bunch of graduate students or postdocs, and a whole bunch of agents working together, which is remarkable.
• "This is going to happen now. This quarter. Not this year. You can see this with the gains in math that have occurred in the last month."

Axios: Are you concerned at all about students getting hired amid an entry-level worker crisis?

• "The graduate students are in heavy demand for all the obvious reasons. Imagine a scientist who can use AI in a transformative way that's graduated in the next few years out of Caltech. This person will get the best jobs everywhere, and it's true in any field.
• "There's a general concern in the world about that. But for what Caltech does, there's zero concern.
• "Caltech is the crown jewel of nerddom."

What they're saying: "Advances in AI and computing hold immense promise not only for accelerating our research enterprise but also for expanding the realms of discovery and exploration," Caltech president Ray Jayawardhana said in a statement to Axios.

• "[Eric] truly understands the significance of this moment and Caltech's vital role in turning possibilities into progress for the benefit of humanity," he added.