Sima Noorani, Shayan Kiyani, Hamed Hassani, George Pappas
The study introduces a framework for human-AI collaboration that ensures AI assists without undermining human strengths, using user-defined rules and an online algorithm to improve decision quality in interactive tasks.
As we increasingly rely on AI to assist in important decision-making, it's crucial to ensure that these systems help rather than hinder us. This research proposes a new way to collaborate with AI by letting users define rules for how AI should complement human decision-making. The framework uses these rules to guide interactions, ensuring AI adds value by compensating for human weaknesses without interfering with their strengths. The approach was tested in medical diagnostics and reasoning tasks, showing that it can effectively improve decision quality by adjusting how strictly the AI follows these rules.
