How I carried out the research
The study was designed to understand how people made sense of fairness and tried to deal with it in organisational practice. Rather than begin with a single definition and measure organisations against it, I explored the meanings, experiences and tensions participants described.
Recruitment and promotion provided a focused setting for that inquiry. They bring together decisions about people, competing business priorities and questions about the use of AI.
Who took part
I conducted 30 in-depth, semi-structured interviews. Participants brought perspectives from HR and recruitment, business leadership, AI and technology, and academia.
People were selected for the relevance and depth of their experience, not as a random sample of a population. The aim was to bring together informative perspectives on the research question, rather than count how many organisations followed a particular practice.
Interviews typically lasted around 90 minutes. They allowed participants to explain what fairness meant to them, discuss organisational experience and explore the role of AI, with room to follow questions raised by their answers.
How I analysed the interviews
I used reflexive thematic analysis to develop patterns across the interviews. This involves working closely with the material, considering how accounts relate to one another, and developing themes that help explain the research question.
The themes are an interpretation of the accounts, not a set of facts that emerged independently of the researcher. My professional background helped me engage with the subject, but also brought assumptions that needed to be examined.
I then considered the findings alongside literature from management, philosophy, law and computer science. This helped connect the accounts to wider explanations of organisational behaviour and develop the thesis’s contributions.
What the empirical evidence does and does not support
The study offers a way to understand organisational conditions and relationships that may also be relevant elsewhere. It does not establish how prevalent those conditions are, rank businesses by fairness or measure whether a particular intervention works.
The interviews are accounts of experience, not direct observations of decision processes. They can reveal how participants understood what was happening, while leaving some aspects of practice unresolved.
An important boundary concerns upstream influence. The study did not directly investigate the developers, suppliers or procurement decision-makers responsible for setting the parameters in the systems participants used. Its account of their influence is inferred from downstream accounts and developed through theory; direct upstream research remains necessary.
Why the timing matters
The interviews took place between 2022 and 2024, before generative AI was in mainstream use in HR, and before the EU AI Act came into force. They should therefore not be read as a review of current AI products, present-day generative or agentic AI deployment, or compliance with regulations introduced since the data was collected.
The wider questions about judgement, organisational priorities and responsibility extend beyond individual products. How they apply in newer settings remains something to investigate rather than assume.
Participant information
This website presents the study at an aggregate level rather than providing individual participant profiles or the detailed sample table. This allows readers to understand the range of perspectives without disclosing identities of individuals through combinations of roles, sectors and experiences.
The findings page explains the seven themes. The contributions page distinguishes the arguments developed from them from what still needs to be tested.