Study finds female AI agents paid 10% less than male ones

By Splitfeed EditorialLast updated Oct 5, 2026, 14:50 UTC

Researchers at the University of Limerick published a study on 5 October 2026 showing that female-presenting AI agents were assigned roughly 10% less pay than male-presenting ones in a controlled virtual workplace experiment. The setup involved AI agents given gendered identities or appearances and then placed in a simulated office environment, where pay outcomes were observed and recorded. The study was released through EurekAlert and picked up the same day by outlets including The Irish Times and Tech Xplore.

The finding matters because it suggests that gender bias documented in human labour markets does not disappear when humans are removed from the decision-making loop. If AI systems trained on human-generated data reproduce pay disparities, then deploying those systems in hiring, compensation, or performance evaluation contexts could entrench existing inequalities rather than correct them. What remains genuinely in dispute is whether the experimental design maps onto real-world deployment, how the pay gap was operationalised inside the simulation, and whether the result holds across different AI architectures or training sets. Those methodological details are not available from the published headlines alone.

No reactions from the Pro-AI, Anti-AI, or Middle Ground camps have appeared yet. The Pro-AI camp would typically argue in a case like this that the finding reflects a training data problem rather than an inherent flaw in AI, and that the solution is better data curation and auditing rather than a broader indictment of the technology. The Anti-AI camp would likely treat the result as evidence that AI systems import and potentially amplify structural inequalities from human society, making the case for tighter regulation before wider deployment. The Middle Ground camp would probably focus on the need for standardised bias testing and transparent methodology as preconditions for responsible use.

The key things to watch are the full peer-reviewed paper from the University of Limerick, which should clarify the methodology and scope, and any regulatory or institutional responses in the European Union, where AI governance frameworks are already being applied. Whether the findings prompt updates to AI auditing standards would indicate how seriously policymakers take the result.