Zach Napora | 2025 I.S. Symposium
狈补尘别:听Zach Napora
罢颈迟濒别:听Man, Machine, or Something in Between? How Emotions and Advisor Identity Shape Our Use of AI Advice
惭补箩辞谤:听Neuroscience
础诲惫颈蝉辞谤:听Grit Herzmann
As artificial intelligence (AI) rapidly integrates into daily life, I鈥檝e always been fascinated by one core question: Why don鈥檛 people trust AI鈥攅ven when it鈥檚 right? My Independent Study explores that question by diving into the psychological factors that influence how people use (or avoid) advice from AI systems. Specifically, I investigated whether emotional states like confidence or anxiety, and the way AI is presented (as a faceless machine versus a friendly, named character), impact how much people rely on AI-generated advice.
In the study, participants first guessed how much money was in jars of coins. Then, after writing about a personal moment of anxiety, confidence, or a neutral experience, they got a second chance鈥攖his time with advice from either a human, an unnamed AI, or an AI with a name, personality, and even a cartoon owl avatar. What I found was surprising: while emotion didn鈥檛 play a major role, people were significantly more likely to follow advice from the anthropomorphized AI than from a human.
This finding suggests that giving AI a relatable identity might be key to increasing trust and improving human-AI collaboration. I鈥檓 excited by the real-world applications, from medical decision aids to educational tools. In the future, I鈥檇 love to explore how far we can push this idea鈥攈ow 鈥渉uman鈥 does an AI need to seem before people really listen? And where鈥檚 the line between helpful and too human?
Posted in Symposium 2025 on May 1, 2025.
