Beyond Blind Trust: When Confident AI Advice Overrides Human Judgment, The Selective Obedience Paradox | By Humna Nadeem

This research examines how confidently presented AI advice affects human decision-making, especially when the AI is wrong. With 50 participants and 250 decisions, it found that people followed AI in 31 cases, including 16 incorrect recommendations. The study highlights the “Selective Obedience Paradox”—people may not blindly trust AI, yet can still follow incorrect advice.