Source record

Published by arXiv cs.HC (T1 source). The original is at https://arxiv.org/abs/2610.00163.

Pipeline notes

The summary and note below are generated by the signal pipeline — they are Beyond Desk’s reading, not quotations from the source.

Two preregistered experiments (N=200, N=198) had participants manage a simulated clothing factory with and without an LLM advisor. AI-supported participants reported higher confidence and less effort, went bankrupt less often, and showed a small knowledge advantage. After AI withdrawal, those who altered recommendations more often performed better unaided.

Preregistered empirical measurement of what remains in the user after the AI is removed, not just in-the-loop performance. Useful framing for teams evaluating AI advisors.

Cited by

No citations on record.