Forced contact moves LLM agent opinions just 2%, polarization study finds
profjamesevans · x · 2026-10-09
From the Evans team's paper thread: rewiring networks could not remove polarization; forcing cross-cutting contact moved opinions about 2%, close to human field experiments; prompting agents to stay open to persuasion cut polarization 11.8%, and neutral influencers cut it 8.8% — in crystallized networks, changing minds beat changing ties. The quoted post notes the testbed's policy value for exploring how interaction rules, network structure, and agent traits combine to polarize, and conditions for mitigation and productive deliberation.
More from Research
- Multi-agent framework CORAL presented at COLM, tops 1.1K GitHub stars — pliang279 · 2026-10-09
- LLM judges of AI-scientist novelty flip on tiny prompt tweaks, AI2 finds — TuhinChakr · 2026-10-09
- TritonRL: an 8B RL-trained model matches 100B+ frontier models on Triton kernel generation — allenainie · 2026-10-09
- LLM agent polarization study from Tsinghua-UChicago team published in Nature Communications — profjamesevans · 2026-10-09
- LLM agents build human-like social networks — and polarize anyway — profjamesevans · 2026-10-09
- Mathematician downplays Navier-Stokes 'mistranslation' uproar: no doubt about the proof — basedjensen · 2026-10-09