Embedding inclusion in events on Responsible AI for Mental Health (RAI4MH)
by Rafael Mestre (project lead), University of Southampton
When we began the Responsible AI for Mental Health (RAI4MH) project, we knew we wanted our workshops to be more than academic gatherings if we really wanted to abide by Responsible Research and Innovation (RRI) principles of reflection and engagement. We all know how to navigate academic workshops, and it is relatively easy to set one up: write some academic blurb, send some emails, open some presentation/poster slots, and academics will come from under the rocks wanting to share their work! Our aim, however, was to understand and discuss the future of responsible AI in mental health, and that requires the perspectives of clinicians, policymakers, the third sector, industry, and people with lived experience. Building that diversity was not easy, but it was essential.
One of the things we already knew, and this experience made it even more clear, was that equality, diversity and inclusion (EDI) does not “just happen.” The pool of participants is already skewed and shaped by systemic imbalances, and if we simply drew on existing networks, we would reproduce those biases and end up with a room of the “usual suspects.” Instead, we had to actively search out people across gender, ethnic background, career stage, and sector. That took much more time than expected. Contacting overstretched people outside academia often meant conversations about why it was worth their while, what they would get out of it, and how we could create a space that wasn’t just another academic talking shop. That also meant that our events had to be carefully designed and not planned last minute.
Accessibility mattered too – removing barriers is one of the key elements of EDI practice. Running the events in hybrid format was complicated and resource-intensive (and much more expensive!), but it made all the difference. At our London workshop, more than 120 people joined online from around the world. That global reach was only possible because of the extra effort to keep the format open and flexible.
Looking back, the main lesson is simple: inclusion requires planning, resources, and persistence. It was still not perfect, but it never is. The audience was still skewed to the academic side, and most definitely certain voices were missing. But we learnt that the extra effort was absolutely worth it – for the richness of debate, the credibility of outputs, and the chances of real impact –, and to be kind to ourselves (and others) when things don’t work out perfectly.