AI and the Dunning-Kruger effect
Generative AI is changing how people access and process information. For many, it’s replacing the need to sift through sources, compare viewpoints and wrestle with ambiguity. Instead, AI delivers clear, polished answers within seconds. While those results may or may not be accurate, they are undeniably efficient. This has already led to big changes in how we work and think.
But this convenience may come at a cost. When people rely on AI to complete tasks and think for them, they may be weakening their ability to think critically, solve complex problems and engage deeply with information. Although research on this point is limited, passively consuming AI-generated content may discourage intellectual curiosity, reduce attention spans and create a dependency that limits long-term cognitive development.
To better understand this risk, consider the Dunning-Kruger effect. This is the phenomenon in which people who are the least knowledgeable and competent tend to be the most confident in their abilities, because they don’t know what they don’t know. In contrast, more competent people tend to be less confident. This is often because they can recognize the complexities they have yet to master.
This framework can be applied to generative AI use. Some users may rely heavily on tools such as ChatGPT to replace their cognitive effort, while others use it to enhance their capabilities. In the former case, they may mistakenly believe they understand a topic because they can repeat AI-generated content. In this way, AI can artificially inflate one’s perceived intelligence while actually reducing cognitive effort.
This creates a divide in how people use AI. Some remain stuck on the “peak of Mount Stupid,” using AI as a substitute for creativity and thinking. Others use it to enhance their existing cognitive capabilities.
In other words, what matters isn’t whether a person uses generative AI, but how. If used uncritically, ChatGPT can lead to intellectual complacency. Users may accept its output without questioning assumptions, seeking alternative viewpoints or conducting deeper analysis. But when used as an aid, it can become a powerful tool for stimulating curiosity, generating ideas, clarifying complex topics and provoking intellectual dialogue.
The difference between ChatGPT making us stupid or enhancing our capabilities rests in how we use it. Generative AI should be used to augment human intelligence, not replace it. That means using ChatGPT to support inquiry, not to shortcut it. It means treating AI responses as the beginning of thought, not the end.
AI, thinking and the future of work
The mass adoption of generative AI, led by the explosive rise of ChatGPT – it reached 100 million users within two months of its release – has, in my view, left internet users at a crossroads. One path leads to intellectual decline: a world where we let AI do the thinking for us. The other offers an opportunity: to expand our brainpower by working in tandem with AI, leveraging its power to enhance our own.
It’s often said that AI won’t take your job, but someone using AI will. But it seems clear to me that people who use AI to replace their own cognitive abilities will be stuck at the peak of Mount Stupid. These AI users will be the easiest to replace.
It’s those who take the augmented approach to AI use who will reach the path of enlightenment, working together with AI to produce results that neither is capable of producing alone. This is where the future of work will eventually go.
This essay started with the question of whether ChatGPT will make us stupid, but I’d like to end with a different question: How will we use ChatGPT to make us smarter?
The answers to both questions depend not on the tool but on users.
Aaron French, Assistant Professor of Information Systems, Kennesaw State University
This article is republished from The Conversation under a Creative Commons license. Read the original article.
