Why do experts miss AI’s errors? Evidence from a randomized labeling experiment
Sofoklis Goulas, Rigissa Megalokonomou, Panagiotis Sotirakopoulos · 2026
Read Why do experts miss AI’s errors? Evidence from a randomized labeling experiment by Sofoklis Goulas, Rigissa Megalokonomou, Panagiotis Sotirakopoulos (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13248211.txt.
Description
Goulas, Megalokonomou, and Sotirakopoulos tested how experts react to wrong scores in a 2026 PNAS Nexus experiment. Education specialists graded the same student work, paired with a score that was deliberately too harsh or too kind. Half the scores were labeled as coming from AI, and half as coming from a person.
When the bad score was harsh, experts moved less toward fairness if they thought AI had set it. When the score was lenient, the label made little difference. The authors trace the gap to how capable and responsible experts judged the algorithm to be, a warning for anyone designing human oversight of automated grading.
Catalog record
| CREATOR | Sofoklis Goulas, Rigissa Megalokonomou, Panagiotis Sotirakopoulos |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 57 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |