The video I didn’t expect to share
Yesterday, Bernie Sanders uploaded this video to his YouTube channel. A direct conversation with Claude, Anthropic’s AI model, about privacy, data, and the impact of artificial intelligence on democracy.
I don’t usually agree with Senator Sanders on political matters. But this video is different. The questions he asks are the right ones. And the answers, coming from the AI itself, have something unsettling about them that’s worth analyzing.
The terms and conditions problem
Sanders asks how much companies know about us and how they get that information. The answer is uncomfortable: browsing history, location, purchases, searches, even how long you pause on a web page. All of that gets combined to build a detailed profile of who you are.
The legal mechanism that makes it possible is one we all know and nobody reads: terms and conditions.
We’re not aware of what we’re agreeing to. And I include myself in that. Terms and conditions are documents so long and deliberately complex that reading them is practically impossible. That’s not an accident. It’s a design choice. If we actually sat down to read what we sign when we create an account on any platform, we either wouldn’t accept it, or we’d accept it with a very different level of discomfort than the automatic click we give it now.
It’s a digital literacy problem. But it’s also a problem of deliberately confusing design. Both things at once.
Microtargeting isn’t just a political problem
Sanders focuses on the political use of data: how AI allows completely different narratives to be shown to different groups of voters, fragmenting shared reality. One voter sees a message about jobs. Another sees one about immigration. Neither knows what the other is seeing.
That’s a real danger. But I think it’s worth pointing out something the video doesn’t fully develop: this isn’t just a political problem. It’s exactly the same mechanism that e-commerce platforms use to show you different prices based on your profile. The same one insurance companies use to calculate premiums. The same one brands use to decide which message hits you hardest based on your detected emotional state.
Microtargeting is a tool. And like any tool, the danger isn’t in one specific use. It’s that it can be applied to anything where someone wants to influence your decision.
Regulation: the car and the road dilemma
Sanders asks whether a moratorium makes sense — an official pause on building new AI data centers — to buy time for regulation. And he points out something I find honest: tech companies are pouring enormous amounts into politics precisely to prevent that regulation from happening.
I find that unethical. No nuance needed.
But there’s a real tension underneath all of this. In a globalized world, a country that regulates alone can fall behind one that doesn’t. If Europe hits the brakes and China doesn’t, what happens? Innovation doesn’t disappear. It relocates.
My position is this: I think AI will follow the same path as the car. First came the automobile. That was the innovation. Then we realized there could be accidents, that it could run people over, that it polluted. And from that came traffic lights, speed limits, seatbelts, emissions standards.
Regulation came after the technology. It almost always does.
The same will happen with AI, almost certainly. The question is whether we’ll get there too late. Whether by the time the rules exist, the damage will already be done and hard to reverse. That’s the part that worries me.
What I take away from the video
Bernie Sanders is right about the diagnosis. Data is being collected without us being aware of what we’re really consenting to. The profile that exists on each of us is far more detailed than we imagine. And that profile is used to influence our decisions, whether to sell us something, make us vote a certain way, or keep us glued to a screen.
I don’t necessarily share all of his solutions. But the questions he raises in this video are exactly the ones we should all be asking ourselves.
The fact that a 84-year-old senator is asking them while talking to an AI on YouTube says quite a lot about the moment we’re in.