Ethan Mollick, an associate professor at Wharton and someone who I consider to be a top AI trend spotter, stated:
Since so much of the AI job debate is among economists, they miss that as jobs are disrupted, professions will compete over new boundaries. Whether there or more or less jobs, they will certainly change, and professions will compete to gain jurisdiction over valuable ones, in large part through regulation, requiring credentials, and appealing to the public.
Sociologists have been thinking about this for a long time, Abbot’s System of Professions is a good read on what happens historically.
I’ll be frank, Abbott’s system of professions is a new theory to me. It describes how expert occupations compete for authority over problems by claiming jurisdiction, using abstract knowledge, and negotiating boundaries within a broader ecology of rival professions.
AI enters that ecology as both a tool and a rival jurisdictional force: it makes parts of diagnosis, inference, and treatment routine, and thus weakens some professional claims while strengthening others that can supervise, interpret, validate, or govern AI-mediated work.
In the modern job market, this means AI does not simply “replace jobs”. Rather, it redistributes jurisdiction by creating new vacancies, forcing occupations to defend what counts as expert judgment, and rewarding groups that can credibly claim authority over AI-enabled workflows.
I think this theory is pretty on point. AI makes jurisdictional dynamics unusually visible, fast, and way more concrete. Even areas with high abstract knowledge is populated by people that can only know so much. I have seen incredible knowledgebase tools built by active professors to extend their personal understanding of their fields of research — I can only surmise that pooling such tooling (especially in a ubiquitous framework) leads to even more extensive knowledge.
Abbott’s theory does appear to lack idiosyncratic market demand (e.g. a regulated industry or government agency may not allow AI workflows internally), but one would expect organizations in such situations would temporarily benefit from a surplus in labor supply capability. In fact, it may be advantageous to delay transitioning while other companies work out the “modernization” wrinkles (link points to UK NCSC on security considerations for AI tooling).