Study finds AI attribution does not change how readers judge gender-biased statements
A Spanish-language experiment using Google autocomplete statements found that telling participants content was AI-generated did not alter their agreement or reaction times.
In plain words
Researchers tested whether people react differently to statements when told some were produced by AI. Spanish-speaking participants rated gender-related statements, some biased ones drawn from Google's autocomplete suggestions and some human-written facts. Knowing the origin made no measurable difference in how quickly or how strongly people agreed.
What was announced
The study, published as an arXiv preprint, recruited Spanish-speaking participants through Prolific and asked them to evaluate two kinds of material: biased statements derived from Google's autocomplete function and human-authored, non-biased gender-related facts. Some participants were told that part of the material was AI-generated, while others received no information about its origin. The authors report that AI attribution affected neither reaction times nor agreement judgments.
The authors interpret this as evidence that AI-generated content does not automatically raise credibility concerns and may even be evaluated with less scrutiny than human-produced content. The finding is limited to the tested setting: gender-related statements in Spanish, with bias measured against autocomplete-derived material, so broader conclusions about other languages, topics or AI systems are not established by this abstract alone.
What it does not show
- The source is only the paper's abstract, so methodological detail, sample size and full results cannot be verified from the supplied material.
- The experiment covered Spanish-speaking participants judging gender-related statements, so generalisation to other languages, domains or AI systems is not supported by the evidence.
Sources
Source-support review completed 17 Nov 2026. This story was drafted with AI assistance and checked against retained source passages. Automated checking is not independent verification.
Coverage: arXiv: abstract.
Updates and corrections
No corrections so far.
First published 17 Nov 2026
New to the subject? AI from the Ground Up starts from the beginning.