There is something quietly disorienting about presenting research on workplace inequality in a country that consistently tops global indices of social trust, gender parity, and worker wellbeing. Kristiansand, the sun-drenched coastal city in southern Norway that hosted European Academy of Management conference (EURAM) this year, is precisely the kind of place that makes you question the universality of your own assumptions which, as it turns out, is exactly the right mood to be in when you are trying to talk about artificial intelligence (AI) and equality in the modern workplace.

I was there to present a paper on how AI and algorithmic systems are reshaping the architecture of human resource management across the UK’s research and innovation sector. The argument is not a comfortable one, and I suspected it would land differently in a Scandinavian room than it might elsewhere. It did.

The paper has its roots in a seed fund I received as a postdoctoral researcher on the EDI Caucus, a UKRI- and British Academy-funded project examining equality, diversity, and inclusion (EDI) across the UK’s research and innovation sector. The topic I chose was to specifically explore the intersection of EDI and algorithmic management. This seed funding each postdoc in the project team received as part of its commitment to researcher development, gave me the space to ask a question that sits slightly outside EDICa’s main survey architecture but runs through its findings like a thread: what happens to equality when the systems governing people’s working lives stop being human and start being automated? What EDICa keeps surfacing, across its data and across every forum where we present, is a tension that is easy to miss if you are looking at AI adoption primarily through a productivity or efficiency lens.

Algorithmic management systems, the scheduling tools, performance monitoring platforms, workload allocation engines, and sentiment analysis dashboards that are steadily embedding themselves into the day-to-day operations of UK research organisations, do not arrive neutral. They are not clean, objective arbiters of fairness. They encode the assumptions of their designers, reflect the historical patterns of the datasets they were trained on, and tend to reproduce the structural inequalities already present in the organisations that deploy them. The people least likely to be visible to these systems — those on fixed-term contracts, disabled researchers, neurodivergent workers, those working reduced or non-standard hours — are often precisely the people whose working patterns fall outside the parameters that algorithmic tools are built to recognise as normal.

This matters because the prevailing narrative around AI in the workplace has been stubbornly optimistic. Technology, the argument goes, removes the subjectivity from human decision-making. It takes the bias out of hiring, the politics out of performance management, the inconsistency out of line management. EDICa’s evidence complicates that story considerably. What we are seeing, instead, is a parallel architecture — a layer of algorithmic governance that sits alongside formal HR policy but is rarely subject to the same scrutiny, equality impact assessment, or accountability structures. It is HR without the paper trail.

The discussion in Kristiansand sharpened something I had been circling for a while. Scandinavian participants were often quicker to reach for a structural framing — the idea that technology governance is fundamentally a labour relations question, not a technical one. That framing felt both more honest and more tractable than the largely managerial lens through which AI adoption tends to be discussed in UK policy contexts. The question is not simply whether an algorithm is biased. It is who decided to deploy it, under what governance framework, and whose working experience it was designed — explicitly or implicitly — to centre.

Diversity, in the changing age of technology, is no longer only a question of representation. It is a question of visibility — of who gets counted, who gets seen, and whose reality the system is built to accommodate. EDICa is taking that argument into academic conferences, policy forums, and parliamentary spaces, because the evidence base needs to be constructed in public, not buried in appendices.

Kristiansand was one stop on that journey. The work continues.

Dr Siddhartha Saxena is a Postdoctoral Research Associate on the EDI Caucus, based at Heriot-Watt University.