Definition. The illusion of competence is the gap between how capable people feel and how capable they are. Fluent material feels learned, a legible explanation feels understood, and a good output feels like proof of a good performer. The gap is invisible from the inside, which is what makes it the central measurement problem in this research.
Confidence and durability point in opposite directions
Bjork and colleagues set out the finding that organises the rest: the study conditions producing the highest confidence produce the least durable learning, and the conditions that feel like they are going badly produce the most. Graded entry. Learners asked to choose their own method reliably choose the one that feels best, which is reliably the weaker one.
Fisher and colleagues then showed the effect can be induced by a tool. Searching the internet inflated people's estimates of their own unaided knowledge, and it did so even on questions the search had never touched. Graded entry. Access to an external source was being experienced as personal knowledge. That study used a search engine in 2015, on a task where the person still had to read, choose and interpret.
The gap that only appears when the tool is taken away
Sankaranarayanan gave 78 participants a programming task in three conditions: manual, unrestricted AI, and a scaffolded version built to make the user do part of the thinking. Both AI groups beat the manual control on the work itself and did not differ from each other. Then the AI was removed and participants had to maintain what they had built. The unrestricted group failed at 77 per cent, against 39 per cent for the scaffolded group. Graded entry.
Two groups, indistinguishable while the tool was present, separated by a factor of two the moment it was not. Bastani's school trial has the same shape: grades up 48 per cent with unrestricted access, then 17 per cent below a control group that never had it, once it was withdrawn. Graded entry.
Huemmer's three-wave longitudinal study puts the metacognitive layer on top, tracking how verification effort declines as confidence in the tool grows. Graded entry.
Why self-report is the wrong instrument
Almost every published survey of AI's effect on skills asks people whether they think their skills have degraded. EY's 2025 survey found 37 per cent worried about it, rising to 43 per cent in the UK. Graded entry. BCG asked 70 executives whether they were observing deskilling and half said yes. Graded entry.
Those are useful figures about worry and about the executive agenda. They are close to worthless as measurements of capability, because the illusion of competence is precisely the claim that people cannot see this in themselves. Asking someone whether AI has degraded their judgement asks them to report on the one thing the effect obscures. The Polish endoscopists whose unassisted detection rate fell six percentage points were not reporting a decline. Somebody measured it. Graded entry.
What defeats it
The illusion is a claim about introspection rather than about performance, so any measurement taken without the assistance defeats it: unaided work samples, live decisions made without the tool, and asking someone to explain and defend an output rather than produce one. That is the argument set out in assessing capability rather than output.
The uncomfortable implication for anyone running an AI programme is that the people best placed to report a problem are the least able to detect it, and the dashboard showing output quality will not show it either. That is why capability debt accrues silently, and why the diagnostic question is not how much AI a team uses but what happens when it is removed.
Key sources
- Bjork, E. L. and Bjork, R. A. (2011). Making Things Hard on Yourself, But in a Good Way: Creating Desirable Difficulties to Enhance Learning. Graded entry.
- Fisher, M., Goddu, M. K. and Keil, F. C. (2015). Searching for Explanations: How the Internet Inflates Estimates of Internal Knowledge. Journal of Experimental Psychology: General. Graded entry.
- Sankaranarayanan, S. (2026). Mitigating "Epistemic Debt" in Generative AI-Scaffolded Novice Programming. Graded entry.
- Budzyn, K. et al. (2025). Endoscopist deskilling risk after exposure to artificial intelligence in colonoscopy. The Lancet Gastroenterology and Hepatology, 10(10), 896-903. Graded entry.
Related SuperSkills research
The illusion of competence is why capability debt is not reported by the people carrying it. The mechanisms it conceals are retrieval practice, productive struggle and desirable difficulty. See also automation complacency, the Google effect, am I becoming dependent on AI, assessing capability rather than output and synthetic seniority.
About this research
Rahim Hirji is the author of SuperSkills (Kogan Page, 2026), keynote speaker on AI and human capability, and founder of The SuperSkills Intelligence Company. The illusion of competence is an established concept in metacognition research and is not a SuperSkills coinage. Findings are attributed to the studies that produced them and kept separate from the interpretation.
Cite this
Hirji, R. (2026). What is the illusion of competence? The SuperSkills Intelligence Company. Last reviewed 30 August 2026. thesuperskills.com/research/what-is-the-illusion-of-competence