Top of Mind: An AI Job Apocalypse? (Goldman Sachs, June 2026): mass job losses unlikely, displacement still real

Goldman Sachs Research's Top of Mind series takes on predictions that AI will wipe out jobs before the end of the decade. MIT's Daron Acemoglu and Neil Thompson and GS economist Joseph Briggs all doubt an apocalypse, but they differ on scale: Briggs expects meaningful but temporary displacement as new jobs emerge, Thompson sees a rising tide rather than a crashing wave, and Acemoglu expects a small net negative over five years with larger losses possible if investment keeps targeting replacement over complementing workers. The issue also includes GS analysis of substitution versus augmentation by occupation, evidence on whether AI is hurting new graduates, the uneven toll across worker groups, the limited labor productivity uplift measured so far, and a cost comparison of AI agents against human workers. The bank's bottom line is that AI's effect on corporate labor needs and earnings is too uncertain for markets to reward yet, which is useful context for founders selling labor-replacement value to enterprise buyers.