AI: does understanding an explanation mean you can explain it?
A manager may have an AI-assisted briefing, find it convincing and still be unable to explain its reasoning to the team. The precise question is therefore: how can we check whether assisted preparation produces usable understanding rather than merely a presentable document?
What the literature establishes
Leonid Rozenblit and Frank C. Keil (2002, Cognitive Science) studied the illusion of explanatory depth: people often believe they understand mechanisms better than they can actually explain them. Asking for a detailed explanation leads them to revise their assessment of their own understanding. This finding distinguishes familiarity with a subject from the ability to reconstruct how it works.
When access to knowledge feels like personal knowledge
Matthew Fisher, Mariel K. Goddu and Frank C. Keil (2015, Journal of Experimental Psychology: General) show that searching for explanations online can inflate estimates of personal knowledge, including on other subjects. The mechanism concerns a blurred boundary between available external resources and the knowledge people possess themselves. For a manager using AI, this suggests a hypothesis worth examining: having an answer available might be mistaken for mastering it.
The premature conclusion
It would be excessive to conclude that AI prevents understanding or that all preparation should be unaided. These studies do not examine generative assistants and cannot directly establish that they produce the observed effects. The useful distinction is narrower: obtaining an explanation and becoming able to explain it are different outcomes. (our executive and employee training programmes)
What a well-written document does not measure
The cited experiments measure neither team leadership nor decision quality in a Lugano business. In practice, assessing only a briefing's clarity leaves open the question of what its author can restate, justify or apply to a different case. Conversely, difficulty speaking without supporting material does not, by itself, demonstrate a lack of understanding: language, speaking confidence and the complexity of the subject also matter.
A practical check in Lugano
As part of SHR — Swiss Human Resources' « Manager à l'ère de l'intelligence artificielle » programme, one exercise involves using AI to prepare an explanation of a fictional, sector-specific case: a banking procedure, a fashion return, a delivery term in trading or delegation in an Italian-speaking family SME. Without consulting the briefing, the manager then explains in Italian the mechanism, one condition under which it holds and a case in which the explanation no longer applies; a subject-matter specialist assesses these elements using a rubric prepared before the exercise. The proportion of elements correctly explained, followed by its change on a comparable case after addressing gaps, provides a verifiable measure of usable understanding without claiming to isolate AI's causal effect. To go further: explore the Managing in the Age of Artificial Intelligence training in Lugano in the canton of Ticino, or browse our executive and employee training programmes in Switzerland.
In pictures: Managing in the Age of Artificial Intelligence in Lugano
- managing in the age of artificial intelligence
- Lugano
- research
- training lugano