When AI Amplifies the Biases of Its Users.

The article focuses on how cognitive biases in human users can amplify biases in generative AI systems, affecting decision-making and outcomes. It explains that bias arises not only from AI training data but also from the dynamic interaction between human thinking and AI outputs, occurring before, during, and after AI prompting. The authors emphasize that awareness of these biases—such as confirmation bias, halo/horns effects, and framing effects—enables individuals and organizations to implement protocols like critical reflection, diverse perspectives, and structured decision-making processes to mitigate bias. By fostering deliberate, critical engagement with AI, leaders can improve the quality of AI-supported decisions and outcomes while reducing the risk of distorted judgments.