Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitchesIn a 2026 virtual-reality office experiment, participants allocated 10.25% less of a shared monetary reward to a female-presenting AI assistant than to a male-presenting assistant that performed the same work. The figure describes money that human participants chose to give each assistant in a controlled setting. It is not a wage that software earned, and it is not a measure of workplace pay gaps.
What participants actually did
The study is a peer-reviewed conference paper titled “Human-Like and Male? How AI Assistant Design Relates to Trust and Monetary Reward at Work in VR.” According to the University of Limerick’s research announcement, it appears in the Proceedings of the 14th Nordic Conference on Human-Computer Interaction (DOI 10.1145/3829807.3829910). The conference was held in Vaasa, Finland, from October 5 to 7, 2026.
As an Amazon Associate I earn from qualifying purchases.
The experiment placed 189 knowledge workers in a virtual-reality office. Each participant completed work tasks alongside AI assistants, rated them, and then divided real money between themselves and an assistant. The academic record describes three assistants whose underlying functionality was identical but whose human-likeness and gender presentation differed. Researchers also interviewed 34 participants to learn how they explained their choices.
Two assistants are named in the announcement: Johan, presented as male, and Johanna, presented as female. Johanna received 10.25% less than Johan for the same work.
#1 Best Overall
The numbers, with their context
| Measure | Reported value | Source | What it covers |
|---|---|---|---|
| Reward allocated to female-presenting Johanna compared with male-presenting Johan | 10.25% less | University of Limerick research announcement, 2026 | Money allocated by participants in the virtual-reality experiment for the same work. HR Dive rounds this to 10%. |
| Participants | 189 knowledge workers | University of Limerick announcement and academic record, 2026 | Workers who completed tasks in the virtual-reality office |
| Interviews | 34 participants | Academic abstract, 2026 | Interviews run alongside the ratings and reward decisions; the abstract does not give the interview method in detail |
| Perceived human-likeness | Female-presenting assistant perceived as less human-like | Academic abstract, 2026 | A directional finding; the abstract does not report its magnitude |
The 10.25% figure should be quoted with its frame attached: it is the gap in an experimental allocation, measured for one set of tasks and one virtual office. When precision matters, use 10.25%. When a rounded figure is enough, say “about 10%” and keep the word “experimental” in the same sentence.
Why “paid” appears in quotation marks
The headline’s quotation marks around “paid” signal that the reward was symbolic in one sense. Participants were not paying wages to a worker with a contract, a salary, or a bank account. They were dividing money in a laboratory-style scenario, and the study measured how that division changed with the assistant’s presentation. Keeping this distinction visible matters, because readers may otherwise assume the finding describes employment practices.
Stated views and measured behavior diverged
The interviews add a second layer. According to the academic abstract, many interviewees said they preferred assistants that were not human-like, and they said gender was irrelevant to them. Their measured ratings and reward allocations nonetheless showed differences by presentation. The abstract does not establish why this gap between what people said and what they did occurred, so readers should treat the mismatch as something the study observed rather than something it explains.
What the result does not establish
- It is not a labor-market pay gap. No employer, wage, or hiring decision was measured.
- It is not a population estimate. The sample was 189 knowledge workers in one experimental setting, and the reported figure applies to that setting.
- It does not isolate gender alone. The design varied both human-likeness and gender presentation, so the two cannot be cleanly separated from the summary alone.
- Its statistics are not fully visible. The announcement and abstract do not report confidence intervals, detailed statistical tables, or the complete task protocol, so the 10.25% figure cannot be placed in a wider statistical context from these sources.
What the researchers said
Dr Mary Hausfeld, assistant professor at the University of Limerick’s Kemmy Business School and a co-author of the study, said: “Such choices may affect how much people trust AI, how they judge its contribution, and even how they financially reward it.” Her statement frames the reward measure as one part of a broader question about trust and perceived contribution, not as a standalone verdict on AI design.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to report this finding accurately
- Name the setting: a virtual-reality office experiment with 189 knowledge workers.
- Describe the outcome as money allocated by human participants, not wages paid to AI.
- Attach the 10.25% figure to the University of Limerick announcement, and note that it reflects a single experiment.
- Mention that the assistants were described as functionally identical, which is what makes the comparison meaningful.
Read this way, the study offers evidence that an AI assistant’s presentation may shape how people trust it and how they reward it in this kind of workplace simulation. It does not show that a female-presenting assistant is generally undervalued at work.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
