Project description
This PhD project explores how organizations translate high-level ideas of “Responsible AI” into the everyday work of product teams building AI-enabled systems. Many companies now adopt ethical guidelines, governance frameworks, or compliance requirements (including those tied to the EU AI Act), yet these ambitions often become decoupled from actual product decisions – remaining symbolic rather than shaping roadmaps, discovery work, design choices, or engineering practices. The project focuses on outcome-driven product management as a key arena where this tension plays out: instead of measuring success by features shipped, product teams are expected to drive behavioural changes in users, customers, internal teams, and business practices, while also meeting ethical and legal expectations.
Using institutional theory – particularly concepts of translation and decoupling – the project investigates how Responsible AI principles travel into organisations, how they are reinterpreted in product cultures and governance structures, and under what conditions they become either integrated or merely performative. Empirically, the research is built on an in-depth single-case study of an AI-intensive software company, supported by more than 30 interviews with product managers, engineers, designers, and governance specialists across the industry to validate and refine a proposed “Sprint Ethics Board” (SEB) model. Data collection includes interviews, documents, process maps, and practice-oriented artefacts such as templates, frameworks, and SEB prototypes.
The overall aim is to develop an interdisciplinary framework that helps organisations meaningfully embed Responsible AI into outcome-driven product management. The project seeks to generate theoretical contributions to organisational research on translation and decoupling, and practical guidance for product leaders, designers, engineers, and compliance teams who want ethics and responsibility to become an integral part of how AI products are imagined, built, and iterated.
Funding
This project is partially funded by Milestone Systems A/S.
Contact
PhD Fellow: Lilit Wecker
Email: lwec@create.aau.dk
Supervisor: Kamal Nasrollahi
Email: kn@create.aau.dk
Co-supervisor: Thomas Ploug
Email: ploug@ikp.aau.dk
Co-supervisor: Jeppe Agger Nielsen
Email: agger@dps.aau.dk
