Dunning-Kruger effect and AI (Artificial Intelligence) 

AI Generated image of multiple colors with different colored blocks

The Dunning-Kruger effect is a cognitive bias where people with limited competence in a particular domain overestimate their abilities [2]. 

This effect was first described by psychologists David Dunning and Justin Kruger in 1999 [2]. They found that those who performed poorly on tests of logic, grammar, and sense of humor often rated their skills far above average [1]. For example, those in the 12th percentile self-rated their expertise to be, on average, in the 62nd percentile [1]. 

The researchers attributed this trend to a problem of metacognition—the ability to analyze one’s own thoughts or performance [1]. “Those with limited knowledge in a domain suffer a dual burden: Not only do they reach mistaken conclusions and make regrettable errors, but their incompetence robs them of the ability to realize it,” they wrote [1]. 

The Dunning-Kruger effect has been found in domains ranging from logical reasoning to emotional intelligence, financial knowledge, and firearm safety [1]. It also applies to people with a solid knowledge base: Individuals rating as high as the 80th percentile for a skill have still been found to overestimate their ability to some degree [1]. 

Inaccurate self-assessment could potentially lead people to making bad decisions, such as choosing a career for which they are unfit or engaging in dangerous behavior [2]. It may also inhibit people from addressing their shortcomings to improve themselves [2]. 

How AI can influence the Dunning-Kruger effect 

One feasible way that AI could influence the Dunning-Kruger effect is by providing feedback and guidance to people who overestimate or underestimate their abilities. For example, an AI system could analyze a person’s performance on a task and compare it with objective criteria or peer benchmarks. Then, the AI system could give the person a realistic assessment of their strengths and weaknesses and suggest ways to improve or use their skills effectively. This could help people overcome their biases and become more aware of their competence levels. 

Another conceivable way that AI could influence the Dunning-Kruger effect is by creating new domains of knowledge and skill that challenge existing human expertise. For example, an AI system could generate novel problems or scenarios that require complex reasoning or creativity. These problems could expose the limitations of human cognition and force people to acknowledge their knowledge gaps or errors. This could also motivate people to learn new things and expand their horizons. Alternatively, an AI system could also demonstrate superior performance or solutions in some domains and inspire people to emulate or collaborate with it. This could foster a growth mindset and a willingness to learn from others. 

These are just some hypothetical examples of how AI could influence the Dunning-Kruger effect. However, the actual impact of AI on human metacognition may depend on factors, such as the design, purpose, and context of the AI system, as well as the personality, motivation, and goals of the human user. Therefore, more research and experimentation are needed to explore the potential benefits and risks of AI for human self-awareness and improvement. 


  1. Dunning–Kruger effect – Wikipedia 
  1. Dunning-Kruger Effect | Psychology Today 
  1. The Dunning-Kruger Effect: What It Is & Why It Matters – Healthline 
  1. The Dunning-Kruger Effect: An Overestimation of Capability – Verywell Mind 

For more on biases, please visit our other articles on Biases and Psychology.