Applying RRI Principles within the AI Equality by Design, Deliberation and Oversight.
written by: Dr. James MacLaren (Operations Lead for AIEbyDDO Project, University of Birmingham)
Our project “AI Equality by Design, Deliberation and Oversight” embeds Equality, Diversity, and Inclusion (EDI) along with the principles of Responsible Research Innovation (RRI), at its core. We are committed to addressing the potential negative impacts of AI systems, such as discrimination and violations of fundamental rights, through active participation in negotiations and deliberations at JTC21 under European Standards Organisations CEN and CENELEC. By contributing to the dialogue and shaping the drafting of European technical standards, we aim to ensure that these principles are reflected in real-world applications of AI.
The four principles of RRI according to the UKRI AREA framework– Anticipate, Reflect, Engage and Act – inform every aspect of our engagement. We instil a commitment to think in advance of the potential impacts that an AI system might have – thinking about who it might impact. We encourage organisations to reflect critically on potential uses and misuses of the proposed systems. We demand that organisations engage with stakeholders, understood in the widest light – including organisational partners, customers, users but also the people impacted by system outputs. Finally, we call for continuous reflection and improvement on the application of the system, throughout its lifecycle – recognising that there are distinct phases of a system – from design to closure.
While we successfully applied the principles of RRI outwardly, integrating them internally has presented challenges. For instance, one is how geographically spread out we are. Our core team has members in Oxford, Birmingham, Yorkshire, Brussels, and Montreal. This makes RRI/EDI focused meetings a challenge and constrains opportunities for informal discussion of issues that arise. Limited shared time—given that most team members work part-time—further restricts opportunities for informal discussions of key issues. Consequently, we often address these matters on an ad hoc basis. For instance, a colleague was conducting a survey of tech developers; we were keen to ensure that it was distributed widely so that all voices could be heard. When we meet in person, we schedule time to discuss RRI/EDI principles and to self-consciously consider them in respect to our own work, not only what we promote to others.
Although these efforts demonstrate our commitment, they also reveal the complexities of practicing what we advocate. This duality highlights both our achievements and the ongoing need for improvement—reminding us that fostering RRI and EDI is as much an internal journey as it is an external mission.