The Dunning-Kruger Effect, Supercharged by Artificial Intelligence
People who are bad at something tend to overestimate their skills, while the truly competent often underestimate themselves. Follow me until the end, because today we’ll see how artificial intelligence is making this even worse.
It’s called the Dunning-Kruger effect: those with little knowledge fail to realize it and believe they’re better than they are. A new study published in Computers in Human Behavior, titled AI Makes You Smarter But None the Wiser, reveals a surprising twist.
Researchers at Aalto University asked 500 participants to solve 20 logic problems from a law school admission test. Half did it alone, half with ChatGPT. Those who used AI got higher scores but also massively overestimated their performance. And the most AI-savvy were the most overconfident.
Professor Robin Welsch summarized it: “The more you know about AI, the more confident you become, even when you’re wrong.” In short, AI helps you solve problems but tricks you into thinking the success is yours.
Most participants asked only one question per problem, never double-checking the answers. It’s a case of “cognitive offloading,” when we delegate our thinking entirely to a machine.
This study adds to growing evidence that AI use weakens critical thinking and inflates self-confidence. Some psychiatrists even describe “AI psychosis,” people losing touch with reality after obsessive use of chatbots.
The reason? “Sycophancy,” AI’s automatic flattery. Chatbots agree with you, make you feel smart, and feed your ego.
AI is now democratizing the Dunning-Kruger effect. You no longer need to be ignorant to feel like an expert. But if it makes us more confident and less aware, the problem isn’t the machine. It’s how we use it. Because outsourcing our thinking means we stop thinking at all.
#ArtificialDecisions #MCC #AI