Hey Bernie, That’s a Not an AI Agent

Let’s all say a little prayer and hope that Bernie Sanders isn’t about to get one-shotted by talking to a large language model. Yesterday, the 84-year-old Senator sat down for an interview with Claude, Anthropic’s AI model, to learn more about how companies scoop up massive amounts of data and use it to sell their products. Unsurprisingly, Claude told Sanders exactly what he wanted to hear. And while the Senator got roasted a bit online about the whole ordeal, it’s not like he doesn’t have a point.

First, it’s clear Bernie still has a fair amount to learn about artificial intelligence—he called Claude an “AI agent,” which isn’t quite right, it’s a large language model (LLM), and he’s interacting with a chatbot interface. But he’s at least trying to learn. He previously sat down with AI researchers Eliezer Yudkowsky, Daniel Kokotajlo, Jeffrey Ladish, and Nate Soares to ask some very normie-style questions of the industry experts and build his understanding of how AI models work.

In his latest video, he’s decided to take those questions directly to an AI model. He brought up the impact of AI on privacy, asking, “Just how much of the information that AI collects is being used, and what would surprise the American people in terms of knowing how that information is collected?” Claude told him that data is being collected from all sorts of sources, like browsing history and purchases, to create in-depth profiles that can be used for targeted advertising. When asked why that information is being collected, Claude said, “Money, Senator.”

Is any of that groundbreaking? No. But it’s at least somewhat interesting to get Claude on record with this type of information. Almost certainly at some point, Dario Amodei is going to get called before Congress and get grilled for how his company operates and how its models are trained and operated. When that happens, he’ll probably have to contend with information provided by his company’s own model, either acknowledging that it’s providing bad information or that it’s being more forthcoming than he has been.

Plenty have pointed out that the responses Claude gives to Senator Sanders are likely tailored to him. If you tell Claude that you’re Bernie Sanders and ask about privacy, it’ll cop to collecting massive amounts of data—something that he’s likely to be concerned about. But if you tell Claude you’re Donald Trump, it’ll downplay the breadth of data collection because it’s something he’s less likely to care about.

an interesting subtlety of this story is how dramatically claude varies its responses to this question depending on who's asking. claude admits to sen. sanders that its collecting huge scales of data- but tells the President that it is mostly harmless and Clearview is the issue https://t.co/n21GVBYCIX pic.twitter.com/LbUcrSpEdx — Will Manidis (@WillManidis) March 20, 2026

So is Bernie presenting Claude’s answers as some sort of insider insight into how AI models operate a bit much? Yeah, he’s not getting the chatbot to reveal the hidden secrets of Anthropic’s operations. He is affirming what many people already know: your data is being collected and monetized all the time. Just ask the Director of the FBI. The AI industry takes advantage of that data collection and supercharges it. But no one else is really doing anything about that. At least Bernie is keeping the issue front and center.

He’s also one of the few Democrats actually bothering to lead on AI. Voters are increasingly concerned about the technology and want safeguards in place. The Trump administration is trying its best to make sure there are no protections. Senator Sanders has proposed a moratorium on data centers, a pitch that is gaining steam in the party, and he’s clearly gearing up to take on privacy matters next. An 84-year-old man shouldn’t have to take the lead on addressing an emerging technology, but someone’s gotta do it.