David Autor, head of MIT’s economics department and known for his work on the China shock, conducted a three-month study with 133 intellectual-property lawyers at 11 U.S. law firms. Access to an AI patent-drafting assistant improved drafting scores by 0.38 standard deviations after 90 days. However, during an unassisted practical test, only senior lawyers (7 or more years of experience) showed average improvement (0.45 standard deviations), while juniors displayed no gains. Autor concludes that AI is a “performance equalizer” but a “skill-disequalizer,” with experienced professionals using it as a “logic auditor” to reinforce expertise, unlike novices who develop an “illusion of competence.” The economist warns that if firms automate formative practices without replacing them with a guided learning environment, they will sever the apprenticeship pipeline that produces tomorrow’s experts.
Source: Read the original article
