aiqbench analysis
This paper establishes formal conditions for human-AI synergy, proving that high algorithmic accuracy can paradoxically inhibit performance if it erodes the human ability to correct errors. The study identifies "Human and Algorithmic Correction" as a necessary mechanism and demonstrates that functional de-correlation between agents allows synergy to persist even amidst cognitive biases like automation bias. These findings offer a rigorous theoretical foundation for designing interaction-centric decision support systems in high-stakes domains such as healthcare.
When Do Humans and Machines Truly Complement Each Other? Formal Conditions and Empirical Evidence for Human-AI Synergy
arXiv:2610.27965v1 Announce Type: new Abstract: Can humans and artificial intelligence cooperate such that their combination yields greater accuracy than either agent alone? This result, termed \emph{synergy} or \emph{complementarity}, is notoriously difficult to achieve, even when both agents possess significant independent accuracy. 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. 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. Beyond its theoretical contribution, the proposed framework provides a foundation for the design and evaluation of knowledge-based decision support systems, where interaction policies must dynamically integrate heterogeneous sources of expertise. We demonstrate this through the analysis of seven clinical case studies. Our empirical results reveal a ``dominance trap,'' where a strong algorithmic advantage can paradoxically inhibit synergy by eroding human corrective intervention. 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. These results provide a formal foundation for key concepts in the human-AI systems literature and offer actionable insights for designing interaction-centric AI systems that foster complementarity.
Abstract:Can humans and artificial intelligence cooperate such that their combination yields greater accuracy than either agent alone? This result, termed \emph{synergy} or \emph{complementarity}, is notoriously difficult to achieve, even when both agents possess significant independent accuracy. 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. 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. Beyond its theoretical contribution, the proposed framework provides a foundation for the design and evaluation of knowledge-based decision support systems, where interaction policies must dynamically integrate heterogeneous sources of expertise. We demonstrate this through the analysis of seven clinical case studies. Our empirical results reveal a ``dominance trap,'' where a strong algorithmic advantage can paradoxically inhibit synergy by eroding human corrective intervention. 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. These results provide a formal foundation for key concepts in the human-AI systems literature and offer actionable insights for designing interaction-centric AI systems that foster complementarity.
This work has been submitted to, and is currently under review at, IEEE Transactions on Human-Machine Systems for possible publication. Copyright may be transferred without notice
Artificial Intelligence (cs.AI)
arXiv:2610.27965 [cs.AI]
(or arXiv:2610.27965v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2610.27965 arXiv-issued DOI via DataCite (pending registration)
Full text under the publisher's open licence · 261 words
Related coverage
On the stability of learning
arXiv · Oct 27, 2026
Stopping and Routing LLM Judge Panels
arXiv · Aug 21, 2026
Global Time-Aware Conditioning (GTC): A Mechanism for Time-Aware Pediatric Sleep Representation Learning
arXiv · Oct 21, 2026
ControlMap: Controllable High-Definition Map Generation for Traffic Scenario Simulation
arXiv · Jun 16, 2026