By Bath Institute for Digital Security and Behaviour members Dr Sachini Weerawardhana and Dr Emma Walker, and guest author Dr Menisha Patel.
Technology is developing at a rapid pace. AI is everywhere, bringing both risks and opportunities for individuals and society. Whilst this fast pace of development means that we can harness the benefits of such technologies as soon as possible, it can also lead to potential issues, challenges, and unexpected consequences being overlooked.
The use of Grok AI to create demeaning images of others following its release is just one example, where the potential harms of such technologies were not adequately considered or addressed before deployment. There is a very current tension between the pace of innovation in the competitive technology landscape and the appropriate consideration of broader impacts and consequences of development.
We already consider such matters through institutional ethics processes. But is this enough given the potentially wide-ranging, pervasive and increasing risks associated with the use of AI? No. As researchers operating in this space, we need to ensure that this consideration continuously remains integrated within the work that we do, rather than being a one-off event. Importantly, we also need to understand the concerns of relevant stakeholders, given the very real life and potentially transformative societal impacts of our research. Therefore, Responsible Research and Innovation (otherwise known as RRI) should very much be at the top of the agenda.
So, here you are. You have run into RRI. The field that directs us towards a broader and more holistic exploration of our innovation: minimising the negative implications through relevant mitigation procedures, to ensure the positive implications can be maximised and felt by society. Importantly, the research process is extended to actively include the voices and expertise of relevant stakeholders. It is hoped that this will create an important alignment between what society desires and our scientific endeavours. Indeed, RRI has become a major focus across the research landscape. All the funders are asking about it on grant applications. You know that it is important and you should do it. But how? What does “doing RRI” look like in reality?
If you are looking for a definitive how-to guide, you may be disappointed as there isn’t one. That’s kind of the point - not that it makes for a more straightforward RRI application or helps you know how best you should integrate it into your work. At least, not at first. However, it is an important and deliberate choice that this field remain non-prescriptive, and for very good reason. It needs to avoid becoming a simple “tick box” activity, as has become the fate of so many compliance-related activities, and instead retain a responsive and proactive “lens”.
This doesn’t mean that there aren’t tools out there to help you. Materials continue to be created and designed to aid consideration and the active application of responsible innovation within the AI research context. Tools range from more conceptual frameworks to tangible prompt cards, for example: The AREA 4P Framework provides a useful grid of probing questions related to the key RRI values and research process. The RAI UK Prompt and Practice Cards use a more tangible approach as a physical card deck (also available online). The cards, informed by the AREA 4P Framework, allow for a fluid yet context-specific consideration of RRI. There are also organisations and repositories dedicated to RRI guidance and support, including the Alan Turing Institute, and Responsible AI Toolkit (Gov.UK)
So, it’s clear that there are several tools and sources of advice out there, but which ones should you use? Where do you even start? Who do you approach? RRI starts with you developing your own project-specific methods and goals. So, it is worth exploring these various approaches and seeing which you prefer and what the best fit is given your project constraints.
Because of this deliberate non-prescriptive approach, meaningful RRI requires commitment, time, and the involvement of the right people. Responsible innovation is vital if we are to harness the benefits of AI effectively across society. As researchers, we have both the privilege and the responsibility, of working in this rapidly evolving field. We therefore need to lead the way in understanding, developing, and integrating responsible innovation practices as a cornerstone of our work. And importantly, we also need to recognise that these approaches can be hugely beneficial to the work that we do, and that we are more likely to develop research outcomes that are acceptable, desirable and ultimately more successful.
Respond
