Media News Ireland is tracking a striking new finding from the technology and workplace beat: even artificial intelligence may not escape old human biases. New research linked to the University of Limerick suggests that when AI assistants are given a female appearance, people may reward them less for doing the exact same work as male-presenting counterparts.
The study adds a sharp new layer to the wider conversation around AI in business, workplace culture and digital design. For employers, developers and HR leaders, it is a reminder that bias may not begin with what an AI system does, but with how it looks, sounds and is framed to users.
Media News Ireland: What the virtual office study found
The research placed 189 workers inside a virtual reality office where they completed job-related tasks with different AI assistants. Some assistants were designed to appear more human-like, while others were presented with male or female attributes.
After the tasks were finished, participants were asked to divide real money between themselves and the AI assistant. That detail matters. Rather than relying only on survey answers, the researchers measured behaviour through an actual financial decision.
The result was hard to ignore:
- Female-looking AI agents received lower monetary rewards than male-presenting agents
- The technology powering the assistants was the same
- Male-presenting assistants were seen as more human-like
- Participants claimed they had no clear gender preference, but their actions suggested otherwise
According to the research, the female-presenting AI assistant was paid 10.25 per cent less than the male-presenting version for equivalent work.
Why this AI workplace bias story matters
This is more than a curious lab result. In News Ireland and across global business circles, companies are rapidly introducing AI tools into everyday work, from scheduling and admin to customer support and internal collaboration. If users carry workplace bias into those interactions, the implications could be significant.
Dr Mary Hausfeld of the University of Limerick’s Kemmy Business School said the findings raise important questions as AI becomes more deeply embedded in professional life. Her core point was clear: people often describe AI as neutral, yet design choices can still trigger the same assumptions seen in human workplaces.
That means businesses cannot treat visual identity as a minor cosmetic issue. In practice, presentation can shape trust, authority and even perceived value.
Key implications for employers and developers
For leaders following Media News, the takeaway is immediate. Organisations deploying AI assistants may need to review not just system performance, but user experience design.
- Bias can transfer: Human stereotypes may carry into AI interactions
- Design influences trust: A more human-like or male-coded assistant may be granted more authority
- Measurement matters: Real-world behaviour can reveal patterns that self-reporting misses
- Governance is essential: AI deployment should include fairness and inclusion checks
Same technology, different treatment
One of the most revealing elements of the study is that the AI capability did not change. The assistants were built on the same underlying technology, yet participants responded differently depending on how the assistant was presented.
That finding cuts to the heart of current debates in Agency News Ireland and corporate innovation coverage. Businesses often focus on model quality, productivity gains and cost savings. But the user-facing layer, including names, voices, avatars and gender cues, may quietly shape outcomes in ways many companies have not fully considered.
The male-presenting assistant in the study received higher rewards and was also viewed as more human-like. That combination is notable because perceived humanity often influences trust, cooperation and status in workplace settings.
A warning for AI product teams
For product designers and enterprise buyers, the lesson is straightforward:
- Do not assume neutrality simply because the code is identical
- Audit the visual and verbal design of workplace AI tools
- Test how users respond across different identity cues
- Build fairness reviews into rollout and procurement processes
In the broader Media Digest around responsible AI, this study stands out because it shows how bias can surface before an algorithm even makes a decision. Sometimes the problem begins with the wrapper, not the engine.
What the study says about the future of work
As AI agents become more common in offices, virtual workplaces and hybrid teams, these findings could shape how organisations approach digital colleagues. Businesses may soon need policy guidance not only on AI accuracy and privacy, but also on representation and behavioural impact.
That makes this an important development in Corporate News Ireland, especially for firms investing in AI-led workflow systems. If staff treat certain AI identities as less capable or less deserving, companies risk reinforcing the same inequality patterns they may already be trying to reduce among human teams.
The study forms part of a broader investigation into how human-like AI affects workplace interaction. Researchers also found that participants tended to trust human-like assistants more and gave them more credit than robotic ones, even when the systems behind them were effectively the same.
In short, appearance still matters, even in digital labour.
The bigger takeaway for Media News Ireland readers
For anyone watching AI adoption through the lens of business, policy or people strategy, this research delivers a simple but powerful message. Technology does not automatically erase bias. In some cases, it may mirror it, package it differently and send it back into the workplace.
That is why Media News Ireland readers should see this study as more than a headline. It is an early warning for employers, software vendors and decision-makers: if AI is going to sit inside everyday work, its design must be treated as a serious organisational issue, not just a branding choice.
The clear takeaway is this: as companies build and deploy workplace AI, fairness must extend beyond algorithms to presentation, perception and human behaviour. In the evolving world of Media News Ireland, this is exactly the kind of detail that will define responsible innovation.
