HBHE Critical Conversations
Narrative Power to Fight the AI Arms Race in MI
April 1, 2026
11:30 am - 12:30 pm
3750 SPH I
1415 Washington Heights
Ann Arbor, MI 48109-2029
Sponsored by: HBHE
Contact Information: Lex Eisenberg alexae@umich.edu
The Speaker: Bryan Smigielski is a campaign organizer leading Sierra Club's data center work in Michigan, focused on utility regulation, energy policy, climate and environmental justice. His organizing practice centers on narrative as a site of multifaceted solidarity and power contestation within systems designed to absorb and dissipate public pressure. Focus of the Discussion: An AI arms race and massive data center expansion premised on the future extraction of trillions of dollars in revenue from global commerce is driving one of the largest buildouts of energy infrastructure in history. This discussion will explore data center organizing and the aesthetics of systemic change. The most effective organizing must construct narrative frames spacious enough to hold together coalitions operating at very different registers of power. Institutional actors and well funded organizations often work under constraints that pull toward incrementalism. Grassroots movements carry moral urgency and a willingness to publicly aim for more radical change. The temptation is to fracture — for institutional actors to retreat into cautious incrementalism and for grassroots movements to abandon them in pursuit of bolder action, splitting apart legal leverage and funded infrastructure from the moral urgency and public energy that could give it force. How might we refuse that polarity? How can we, as Joanna Macy writes, "not limit our choices to the outcomes that seem likely, but focus on what we truly and deeply long for, and take determined steps in that direction"? Drawing from the "No Secret Deals" data center campaign in Michigan, we'll examine how organizers built a narrative architecture that housed these tensions, and what a successful movement to fight back against unhinged AI buildout could look like.