Recent research indicates that the widespread use of artificial intelligence (AI) can create a false sense of competence, significantly altering the classic psychological phenomenon known as the Dunning-Kruger effect.

Traditionally, this effect describes a cognitive bias in which individuals with limited skills tend to overestimate their abilities, while those with high expertise often underestimate themselves.

The new study conducted by Aalto University, in collaboration with researchers in Germany and Canada, suggests that engaging with AI flattens this curve, causing users across all skill levels to overrate their performance.

In the experiment, 500 participants completed logical reasoning tasks from the Law School Admission Test, with half allowed to use AI chatbots. Participants were subsequently asked to assess both their task performance and their AI literacy.

Results revealed that all AI users, regardless of prior proficiency, placed excessive trust in the solutions provided. Those more experienced with AI were particularly susceptible to overconfidence, likely because they engaged in "cognitive offloading," accepting answers quickly without thorough reflection or verification.

This reduced engagement in metacognitive monitoring, the process by which individuals evaluate their own reasoning and accuracy.

The study highlights a crucial trade-off: although AI can enhance overall task performance, it simultaneously diminishes the user's capacity to judge their own competence accurately.

As AI becomes increasingly prevalent, this phenomenon could encourage overestimation of abilities, particularly among highly AI-literate users, potentially contributing to flawed decision-making and erosion of critical thinking skills. To mitigate these risks, the researchers recommend designing AI systems that actively promote reflection, prompting users to question their confidence, consider alternative solutions, and engage in more deliberate reasoning.

This approach seeks to balance technical proficiency with metacognitive awareness and critical evaluation, emphasizing the need to integrate thoughtful assessment into AI interaction.