Study: Social media probably can't be fixed
Study: Social media probably can't be fixed
This article reports on a study by Petter Törnberg and Maik Larooij, which concludes that social media's core dysfunctions—including polarization, echo chambers, and the amplification of extreme views—are deeply embedded in its structural design. The research, utilizing a hybrid model of agent-based modeling and large language models, found that these negative dynamics naturally emerge from basic social network architecture, making conventional interventions largely ineffective. The findings suggest that only a fundamental redesign could address these persistent problems.
Note
I normally don't care for reading q/a, but I don't mind it when it's about authors of a research paper.
it's better than reading university press summaries of new research.
using LLMs to study societal patterns seem both incorrect but probably better than what other tools or studies could do (in such a short time frame).
social media definitely seems like a technology with more downsides than good (for society, maybe not so much on the individual level). but the discussion about how this permeates to news and private conversations is a bit of a downer.
I do think there are more ideas worth testing though. it's unclear that the conclusion that the underlying flaw of social media infrastructure offers us no room for solutions.
but I'm certain they won't look anything like we have today. profit motives and even social motives may all have to change .
Published
Summary
Image
