MIT Study Finds AI Creates Illusion of Competence
David Autor warns that AI improves immediate performance for junior professionals without increasing their underlying skills, potentially damaging the professional apprenticeship pipeline.
Research led by David Autor, head of the economics department at the Massachusetts Institute of Technology, indicates that AI serves as a performance equalizer but a skill-disequalizer for early-career professionals. In a three-month experiment involving 133 patent lawyers across 11 U.S. firms, researchers found that AI improved the quality of initial drafts for all participants. However, only senior lawyers with at least seven years of experience demonstrated improved independent judgment after the tool was removed.
Junior lawyers showed no average gain in skill, leading Autor to warn against an "illusion of competence" where novices mistake AI-generated output for their own expertise. The study, published as a National Bureau of Economic Research working paper and conducted by Google, suggests that experienced practitioners use AI as a "logic auditor" to refine strategic scope. In contrast, junior staff may use the technology as a cushion that bypasses the formative practice required to master their field.
Autor cautioned employers that automating entry-level tasks without implementing guided learning could sever the apprenticeship pipeline. He noted that if AI is viewed as a replacement for mastering expertise, the professional community will be sorely disappointed.
