Yearly Research Retreat with AI Policy Lab @Umeå University and the Responsible AI group

Dates: 17-20 March 2026
Format: Responsible AI Retreat

Just back from our yearly research retreat with AI Policy Lab @Umeå University, the Responsible AI group at Department of Computing Science and colleagues from different places.

Our theme this year was Situated AI: grounding AI research in place, community, and lived knowledge rather than a view from nowhere that mascarades as objectivity.
We talked about solarpunk visions for AI at community scale: whose resilience, whose future, built on whose knowledge? We sat with the uncomfortable truth that participation can be co-opted, that inviting more voices into a process doesn’t redistribute power, and can even become a new form of data extraction.

And we turned the lens on ourselves. The publish-or-perish pressure of academia doesn’t just shape what gets said, it shapes who gets to say it, and on what timeline. The incentive structures of academic AI research can reproduce the very dynamics we critique from the outside.

Our working conclusion, borrowed from Donna Haraway: “stay with the trouble”. Sometimes not resolving tensions prematurely, but staying in them long enough, is the most honest thing we can do.

Next up: figuring out what situated AI looks like not just as critique, but as practice, that is, research that is itself accountable to the communities it studies.