Aired Dec 10, 2025

We’ve spent more than a decade blaming the algorithm for toxic polarization — the outrage, the tribalism, the misinformation. If we could just tweak the recommendation system… maybe we could fix what’s broken.

But what if the real problem goes deeper?

In this episode, researcher Petter Törnberg explains why social media polarization is baked into the architecture of platforms — even without algorithm-driven feeds. And now, into that already fragile digital public sphere, comes something far more powerful:

AI that can influence what we think, without us noticing.

From generative AI chatbots to personalized persuasion systems, Petter reveals how this next wave of technology could reshape politics, public opinion, and democracy — in ways that even social media never could.

“The battlefield of the coming years is how AI becomes a form of political power to shape our interests and our worldviews, in a way that is even more profound and dangerous than social media was to our politics and society.”

– Petter Törnberg

If you’ve ever wondered how much control we really have over our beliefs in a world driven by attention and engagement… this conversation is going to change how you see the future of digital life.

Key Discussion Points

• Why fixing the algorithm won’t fix social media
• How platforms create outrage and polarization by design
• What happens when AI starts shaping our interests and values
• The hidden power imbalance between users and tech companies
• Why the future of political persuasion is automated
• What we can do to protect autonomy and democracy

Links

Petter Törnberg (pettertornberg.com)

Can We Fix Social Media? Testing Prosocial Interventions using Generative Social Simulation (Larooij & Törnberg, 2025)

Ars Technica Interview: Social media probably can’t be fixed