Bots, Outrage, and the Myth of the “Fixable” Feed

Bots, Outrage, and the Myth of the “Fixable” Feed

Social media platforms aren’t broken by accident, they reward speed, outrage, and concentration. A bot-only social network proved it: six ‘prosocial’ fixes failed or backfired. Still think a UI tweak will save us?

In a recent study, researchers built an entire social network where every account was a bot, and it still devolved into echo chambers, outrage amplification, and winner-take-all attention. Maik Larooij and Petter Törnberg from the University of Amsterdam used GPT-4o agents to post, repost, and follow while testing six “prosocial” fixes. The results: tweaks helped a little, but sometimes made things worse.

Chronological feeds flattened attention, yet surfaced more extreme content. Boosting opposite views barely bridged divides. “Bridging attributes” promoted civility, while concentrating reach among a narrow set of voices. Hiding follower counts or bios did almost nothing.

The underlying engine stayed the same: reactive engagement builds the network that then feeds the next round of reactive engagement. That loop is the product. In my new book Now What? How to Ride the Tsunami of Change, I argue for a change in perspective, not just a change in interfaces.

Here’s what I’d do next:

• Redefine success around constructive exchanges, not clicks.
• Break repost virality in political threads for cooling, not quieting.
• Open APIs for audited “public-interest rankers.”

Guardrails beat wishful thinking. The study shows that echo chambers aren’t an accident, they’re the business model. If tweaks won’t fix social media, should we redesign platforms entirely, or is it time to build alternatives from scratch?

Read the full article on Futurism...

----