Workers who know their boss is going to review them are actually *more* likely to just mindlessly copy and paste from an AI.
While managers assume oversight encourages scrutiny, the fear of evaluation actually makes workers use AI as a 'shield.' By deferring to the AI's perceived expertise, employees avoid the personal risk of introducing their own errors, effectively outsourcing their accountability to the machine.
Review Makes Workers Less Likely to Revise AI Output
SSRN · 6325399
Can review-a cornerstone of organizational accountability-encourage workers to collaborate with AI rather than outsource their work entirely to it? Across seven experiments (N = 2,895), we find that review has the opposite effect: participants who expected their work to be evaluated by a peer or supervisor modified AI-generated content significantly less than those who did not anticipate review. In other words, reviewing a worker's final work makes them more likely to copy and paste from AI rath