3 Questions you’ve never dared to ask

“Every employee tends to rise to their level of incompetence.” How can we consider the possible disappearance of in-house experts without invoking the Peter Principle? What if AI were to challenge that?

By Kevin Erkeletyan

HAS AI RAISED US ALL TO OUR LEVEL OF INCOMPETENCE?

In 1970, Canadian educationalist Laurence E. Peter put forward an idea based on years of observation: “Every employee tends to rise to their level of incompetence.” The Peter Principle describes a simple mechanism: an employee who is competent in their field (let’s call them an “expert”) is usually promoted until they reach a position in which they are no longer competent. They remain there forever, unable to rise any higher but sure of maintaining their position. They then become a dead weight and hold the hierarchy back.

The Peter Principle has since been the subject of research, including a study by Alan Benson (University of Minnesota), Danielle Li (MIT) and Kelly Shue (Yale), who in 2018, verified some of his claims. But has the emergence of generative AI, in late 2022, changed the game? At first glance, AI hasn’t made us any less competent: we haven’t suddenly become less capable of performing such-and-such a task. But has it nullified or neutralised our competence? If AI can perform a task just as well as we can; if its ability to do it is the same as ours, then it is a zero-sum game: its ability minus ours equals zero. So AI does not accelerate our rise to our level of incompetence, but perhaps in some cases it makes us reach our level of indifference.

CAN AI REACH OUR THRESHOLD OF INDIFFERENCE?

We sometimes forget that AI is human, all too human. The poor thing is therefore subject to the Peter Principle. Any AI tends to rise to its level of incompetence. As it meets our demands, we push it to its limits. In this way, it climbs the ranks of our hierarchy until it is unable to meet our ever-increasing requirements. But when AI becomes incompetent, maybe it is we who become competent again.

CAN AI FREE US FROM OUR LEVEL OF INCOMPETENCE?

While generative AI may make us reach our threshold of indifference towards it (see above), it raises questions about our ability to reinvent our expertise: how can we push against this level of indifference? How can we revive the difference? How can we renew our expertise? AI pushes us to become experts in a different way, and this often involves developing expertise in AI itself. An area of expertise – or rather, our own – means reinventing our competence, adapting AI to our activity and making it the specific lever of our capabilities. For a graphic designer, it means re-examining their creative process through Midjourney; for a translator, training Gemini to recognise their style.

AI spurs us to develop and thus to learn: it can therefore remove the personal limitation that previously brought us to our level of incompetence. Let’s take an example used by Laurence E. Peter. A very good mechanic, whose results are better than those of his colleagues, will probably be promoted to foreman. At this point, he will reach his level of incompetence because he is a poor organiser. But if he goes about it the right way, AI can do that particular job for him. His level of incompetence will then rise to a new threshold and he will again be in line for promotion. So, maybe AI can neutralise this Peter principle. Unless it just pushes the level of incompetence higher. If AI can reinvent our competence, it can surely do the same with our incompetence. The Peter Principle has a bright future ahead…