A new pilot study published on arXiv suggests that human capital, rather than model benchmarks, is the key predictor of success in hybrid human-AI forecasting. Analyzing data from the Polymarket prediction market, researchers found that most individuals either matched the AI's performance or performed worse when collaborating with AI. However, a minority of forecasters demonstrated genuine complementary reasoning, achieving accuracy comparable to or better than the market itself. This elite group was distinguished by collaborative traits like perspective-taking and intellectual humility, rather than raw cognitive ability. AI
IMPACT Highlights the importance of human traits over AI capabilities for effective human-AI collaboration in forecasting tasks.
RANK_REASON Research paper published on arXiv. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
