Researchers at the University of Limerick have made a striking discovery, proving that people tend to project gender bias even onto artificial intelligence systems. During a large-scale scientific experiment, it was recorded that people systematically paid virtual avatars with female appearances roughly 10 percent less for completed work than identical male counterparts. This study was published on the prestigious ACM Digital Library platform and has already sparked widespread resonance in academic and technological circles.
Experiment Essence and Methodology
To test their hypotheses, the scientists involved 189 participants engaged in a detailed virtual reality workplace simulation. The main objective of the experiment was to track changes in human attitudes toward AI assistants depending on their gender visualization. Crucially, the technical foundation, functionality, and task performance speed of all digital assistants remained entirely identical.
Financial Discrimination in Figures
Upon completing the simulation, participants received real money, which they had to distribute fairly among themselves and the working AI agents. Despite the equal contribution of algorithms to the overall outcome, female avatars received significantly lower financial compensation. Dr. Mary Gausfeld emphasizes that giving a digital assistant external features of a specific gender instantly activates deeply rooted human stereotypes.
Social Consequences and Industry Takeaways
The obtained results force a completely new perspective on designing modern human-like interfaces. Developers can no longer view the voice, appearance, or personality of AI as mere harmless decoration. As autonomous agents become actively integrated into modern workplaces, there is a tremendous risk that old social inequalities will simply be reproduced in a new digital form.
