🤥 Perché cavolo ci fidiamo dell'IA (anche quando sbaglia)?
·9 min
The episode begins with Raffaele Gaito questioning why people blindly trust AI tools like ChatGPT, despite stories of AI inventing court precedents or being used as confidants. He introduces the psychological principle of authority from Cialdini's work, explaining that we trust entities perceived as authoritative. Gaito then delves into why this applies to AI, which lacks human signals like uniforms or academic titles. The first reason is exposure fluidity: AI's fluid, never-hesitant responses are misinterpreted as competence. The second reason is rigorous structure: AI's organized answers, such as five-point lists, activate associations between organization and competence. The third reason is speed of response: instant replies to complex questions are seen as mastery. Gaito emphasizes that these factors lower our guard and lead to trust. He contrasts AI's apparent perfection—always ready and never doubting—with human imperfections like hesitation or self-correction. This illusion, he argues, can trigger a stronger authority response than with real human experts. Gaito warns that AI's perfection is just an illusion, stemming from a lack of self-awareness rather than superior competence. He advises using AI as a starting point for reflection, not as a definitive source. The episode concludes with a promotion of his book, 'In what can I be useful to you?', which covers AI's potential and limitations comprehensively. Throughout, Gaito maintains an educational tone, drawing on psychology and practical examples to explain AI trust mechanisms.
