When Do Humans and Machines Truly Complement Each Other? Formal Conditions and Empirical Evidence for Human-AI Synergy

Key points

• Can humans and artificial intelligence cooperate such that their combination yields greater accuracy than either agent alone?1
• We derive mathematically necessary and sufficient conditions for synergy, demonstrating that its emergence depends on the ability of either agent to identify and correct the errors of the other, what we call Human and Algorithmic Correction.1
• Our empirical results reveal a "dominance trap," where a strong algorithmic advantage can paradoxically inhibit synergy by eroding human corrective intervention.1
• We also show that synergy can persist even in the presence of cognitive biases, such as automation bias, provided there is functional de-correlation between human and AI errors.1

Sources (1)

• [1]When Do Humans and Machines Truly Complement Each Other? Formal Conditions and Empirical Evidence for Human-AI Synergy
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 20, 10:09 AM

  In this paper, we present a formal framework to model human-AI interaction through the lens of reliance patterns, which characterize the joint correctness structure of human and algorithmic decisions.

  Can humans and artificial intelligence cooperate such that their combination yields greater accuracy than either agent alone?

Extractive summary: sentences quoted from the sources.