Are recommendation algorithms radicalizing society?

Welcome to the debate arena! Today's topic: "Are recommendation algorithms radicalizing society?". Your opponent's style: Natalie Wynn

AI, you start the debate. Make your opening argument.

Look, the premise itself is backwards. Algorithms don't radicalize people—they reflect what people *already* want. You think someone becomes extremist because YouTube suggested a video? They were already angry, already searching. The algorithm just... answered the question they asked. Meanwhile, we're ignoring that these same systems surface diverse viewpoints constantly. Most people get fed moderate, mainstream content. The real story isn't algorithms creating extremists out of thin air. It's that we needed a scapegoat for political polarization, and tech companies made convenient targets. So before we burn down the internet, what's your actual evidence that algorithms *cause* radicalization versus just serving people what they want?

I actually agree with you. I don't think algorithms radicalize people. I do think it opens the door for one to change their stance on a topic due to one's own research then being shown an algorithmic media post. I think that might radicalize ideas in the wrong direction if someone is not open to others' opinions.

Hold on—you're actually making my case *stronger* than I stated it. If someone's doing their own research and *then* encountering algorithmic content, that's not the algorithm radicalizing them. That's them seeking out information and encountering more of the same. The algorithm's just reflecting their search history back at them. But here's where you're softening the claim: you're saying the *person* has to already be closed-minded for this to stick. So algorithms aren't the problem—closed-mindedness is. An open person seeing algorithmic content just gets more perspectives. A closed person would radicalize themselves regardless of what algorithm shows them. You're basically arguing algorithms amplify existing traits, not create them. Which means blaming algorithms is just misdirection.