GARE Public Resources

GARE Conversation: AI, Data Centers, and the Myth of Inevitability (8.6.26) 

08-27-2026 06:35 PM

Session Description:

In this GARE Conversation, we delved into the impacts of Artificial Intelligence on jurisdictions, community pushback against data centers, and promoting narratives that don’t perpetuate messages of inevitability as we seek to identify solutions and strategies that center the communities put most at risk by this technology.

 
Session Leads:

  • Mandy Mitchell (she/her), Fmr. Director of Policy and Data, Department of Civil Rights, Detroit, MI

    • I grew up "behind the Orange Curtain" in Orange County, California, surrounded by religious and political conservatism. I earned my BA, then MA and PhD, in philosophy, specializing in political philosophy, and taught in Southern California colleges for 15 years. I returned to school for a Master's in Policy just before the pandemic, and found employment with the City of Detroit after graduating.

  • Alejandro Navarrete (he/him), Policy and Research Director, Detroit Action

    • Alejandro Navarrete was raised in and currently lives in Southwest Detroit. He has worked in economic development and policy advocacy organizations throughout SE Michigan. Today, he serves as the Research and Policy Director for Detroit Action, a grassroots advocacy organization . He holds a B.A. in Economics from the University of Michigan and an M.S. in Computational Analysis and Public Policy from the University of Chicago.

  • Theo Pride (he/him), Organizing & Operations Manager, Detroit People’s Platform

    • Theo Pride is the Organizing & Operations Manager for Detroit People’s Platform (DPP), a community-based organization dedicated to building political power with Black working-class Detroiters. His work focuses on challenging the use of public resources for private, profit-driven development and advancing policy strategies that support equitable development and a solidarity economy.

 
When: Thursday, August 6, 2026


Pre-Share Materials:



Shared Materials


Q&A:

  • I haven't looked, but am interested to learn what the AI and Data Center matrix-building looks like in other countries. Any recommendations on where to learn more about that?
    • I’m not really versed in this. I know that the EU has privacy protections that are far more robust than those in the U.S., which are basically limited to HIPAA and FERPA. See here for the GDPR. It does look like the EU Commission has passed an AI Act, but I’m not sure how robust it is. [Mandy]
    • This site shows the EU regulatory framework and links to the act that Mandy referenced.
  • (comment) In some places, voting season is approaching. May want to consider what role, if any, this may play with up and coming electeds and their platforms.
    • Michigan recently held primaries, and data centers were definitely an issue. Ann Arbor’s mayoral candidate is strongly opposed to data centers (there is one going being built in neighboring Saline), but candidates for Senate and Governor are only calling for regulation and transparency rather than prohibition. I don’t see anyone publicly calling the AI phenomenon a bubble or discussing that data centers have a short lifespan (2-5 years). [Mandy]

  • Engaging in deep discourse about AI and LLM can be very polarizing and emotionally triggering. I guess because of the many moral areas to look at the issue. From environmental impact, to the inevitability argument, to assumptions around plagiarism. What would you recommend to frame early conversations? 
    • I agree–it’s a tough issue because people who are using AI will likely feel criticized. I would love to hear how others navigate this, but I usually try to open by making it clear that I understand that ultimately, responsibility lies at the systems/institution level, i.e. with Big Tech, with Congress, with employers’ procurement and management decisions, and that the decisions made at those levels constrain our possibilities for action at the individual level. I have also (in a staff meeting) raised that I have ethical issues with AI, particularly with the equity, intellectual property, and de-skilling aspects of it. I think that lots of people are actually quite concerned about de-skilling (AI making us dumber), which can be a way to connect to people about it. Similarly, I’ve found that it’s easier to establish solidarity around opposition to one’s work being stolen. I think, also, there’s common concern about utility bills going up, which can be a way into the equity conversation. [Mandy]
  • Is there a means to have the benefits (medical breakthroughs) of AI without the development of Data Center.
    • I’m not an expert on this, but I believe the answer is yes. Cal Newport has argued (summarizing Yann LeCun) that rather than ‘general’ AI, which aims to put the whole text-based internet into a training set, the future is ‘modular’ AI, which is domain-specific and requires fewer inputs and far less computing power. Basically, this is AI that would live on a chip rather than needing a data center to run. So you could have a domain-specific AI just for medicine, or for various fields of medicine. [Mandy] 

  • Is anybody else dealing with disability access being used as a defense of AI?
    • This resonated with several practitioners who are navigating this
    • Social justice is intersectional. Pitting marginalized communities against each other to defend AI is disingenuous. These discussions require nuance and AI is not inherently beneficial for disabled people.
    • How Disabled People Are Using AI in Everyday Life (10 page report detailing mixed results)
      • ^^ “The rapid advancement of Artificial Intelligence (AI) presents a similar dual reality: it has the potential to become one of the most transformative accessibility technologies in modern history, but only if people with disabilities are included throughout its design, testing and deployment.“
    • I like to compare ChatGPT use to single-use plastics use–in most cases, convenient, but unnecessary. For some persons with disabilities, however, plastic straws are pretty necessary. I think also about how I’d love to get 1:1 computer use out of my kids’ classrooms, but that for some kids with disabilities, having a laptop for voice-to-text is a necessary accommodation. I’m not sure what the answer is, but I think we can fall back on two principles: (1) public engagement with many groups is necessary in policymaking. That’s just not happening right now. Instead, Big Tech is using everyone as guinea pigs, and pushing governments and other employers to adopt early and often based largely on vibes. (2) the media tends to (likes to) discuss these issues using a “this team vs. that team” framework, but the reality is that there are often many more options available to us. “All the AI, including for people with disabilities” vs. “No AI, people with disabilities lack essential accommodations” runs afoul of the “false dilemma” fallacy. [Mandy]

  • I definitely need talking points to oppose AI in the workplace. I am morally and ethically opposed to the use of AI and building of data centers in my personal life and it is getting harder to avoid using it at work.
    • This is so valid and I hope we have more opportunities for learning share. I want to acknowledge that this is an emerging issue, that none of us really have satisfactory answers, and that the economic and political pressures on managers are intense. I would point to what I said above–bringing in the systems/institutional-level considerations can make conversations less threatening, as can finding easy common ground (utility bills, climate impact, de-skilling, privacy/intellectual property violations, equity) when moving to discuss ethical reservations at the individual level. Not to be wishy-washy, but I think each person has to weigh both their own responsibility to speak up/act AND their power/influence/vulnerability in the workplace. [Mandy]
  • In my city, there is a surprising eagerness to adopt AI tools across the city, and when I raise concerns about conflicts of these tools with our equity and environmental commitments, the promise of productivity increases or decreased workload seems to win out over anything else, are there any tools or resources to address this?
      • You’re up against AI marketing, Big Tech values (convenience over all other goods), capitalist/evolutionary values (increased productivity over all other goods, endless growth), and anti-statism (efforts to shrink government). Hard to counter in a conversation. 
  • (rhetorical, but may generate thoughts/responses) I feel like the point about conservative AI vs. innovation and creativity is super interesting - it makes me think of the question - why are we putting this emphasis on efficiency and speed at all? Is quicker always better?
      • I love this question because it goes straight to underlying values/highest goals. Ultimately, there’s not an answer to this question in the same sense that there’s an answer to a math problem, which can seem unsatisfying, but we wouldn’t be human if we didn’t raise these questions and want to discuss the trade-offs that accompany pursuing different final goals. 

  • How do we communicate the risks and ethical concerns about AI when people are buying into the hype/concerned about being left behind? Are there model ethics policies that help government practitioners evaluate new AI platforms more thoroughly?
      • As I said in another response, we’re up against a Big Tech fantasy, both in the narrow sense of this technology, and in the broader sense of the mythology the entire sector has built around itself. It’s quite hard to break through that. I tend to fall back on a few principles that I think should inform technology adoption:

      1. The economy was made to serve people, not people to serve the economy. When I taught Business Ethics, I always used to hear my students say some version of “that’s not the way the world works”. We’re accustomed to thinking of the economy in this way–we have to adjust and adapt to fit into it as if it’s a “fact” in the same way the natural world is. The truth is that the economy is social. It’s created by many, many decisions, it’s complex, it has emergent properties, but it is a social artifact. We made it what it is to meet our material needs, and we can change it to meet our material needs in better (more ethical, more environmentally friendly) ways. I see this commitment as a version of humanism. 

      2. Adopt tech “at the speed of trust”. I think there’s a business school/management way to think about this phrase, but I think of it as a standard of broad, equitable (anti-racist, anti-classist, anti-ableist, etc.) public engagement with a goal of consensus. 

      3. The precautionary principle. Once upon a time, “buyer beware” was the rule in consumer purchases. If you bought a shoddy product, then it was your fault for not doing your homework. Ralph Nader and others were part of a movement to institute more robust consumer protections, especially in the auto industry, recognizing that consumers don’t have the expertise or time to determine the quality of many goods. Now the manufacturer takes a lot more responsibility to ensure product safety and quality. 

      The precautionary principle does a similar burden or risk-shifting. The environment should not bear the burden of polluting technologies in the name of innovation. Consumer health should not be put at risk by new products made with relatively untested chemicals. Similarly, before we adopt a new technology, we should require corporations to guarantee quality and safety, in the broadest of senses. [Mandy]

Chat Notes & Comments:

  • Large Language Model - a type of A.I program trained on massive amounts of text to recognize patterns, write code, and generate human-like responses.

  • Large Language Model. This is not the only approach to AI, and it's not even a very effective one for most applications, but it is the water/energy-hogging strain we have become infatuated with these past 3 years.

  • The impacts of AI on education at all levels, with students increasingly not putting in even the minimal effort for assignments, is deeply troubling. This is such a helpful framework! (referring to Mandy’s portion of the presentation)

  • Detroit has a better understanding of what it means to turn over their economy to one industry than probably anywhere else in the country. For all the faults of the auto industry of the early 20th century, at least those factories needed people. Hundreds of thousands of workers took those jobs, formed unions, and built good lives. Data centers bring all the harms with none of the benefits for working people.

  • My agency has a policy. We handle sensitive data and have legal and ethical requirements around confidentiality. The policy is restrictive enough to forbid most uses of AI, without separate approval 

  • It's also concerning that so much software forces you to use AI whether you want to or not. It's becoming either you use AI or you don't use any technology at all

  • There's quite a discussion within the language access profession regarding the use of AI and challenges with conversions. 

  • RE Construction jobs: well there is a labor shortage, but if the gov wants construction jobs, there are so many transportation and infrastructure projects that need to be taken on nationwide (failing roads, no public transport, old pipes, etc).. Would that be developing the economy? Maybe not, but at least then we would have better infrastructure

  • I was at the commencement for the University of Arizona and have never seen so many people from different age, race, gender, cultural, class, etc. backgrounds all boo someone like they did for the former google CEO. a very broad coalition 

  • Yes to AI pressure in the office.  I work for city government, and there is interest from management to use google translate rather than human translators because of cost.  In addition, our tech folks are rolling out copilot chat and encouraging folks to use it as the city lays off dozens of people.  It's hard not to feel like these things are connected. 

    • ^+1 on navigating pressures in the office, for sure! Maybe a subtopic or branch, but would love conversations on differential impacts by employee positioning (more junior/early-career employees, gender dynamics, etc.)

  • Working in public health, this is such an important topic in their state and agency. There’s a rush to use AI tools in Washington state, and their union is meeting to discuss the potential policies (which are presented in a very technocratic way). There are a lot of climate concerns in the state, especially for Indigenous Nations in the state, and plenty of concern around data privacy and safety

  • There is a lot of work that needs to happen in the heart of data center country here in Northern Virginia. There is definitely a lot of pressure to push AI use in the office but several staff, myself included, are determined to put some use guidelines in place here. This conversation was a helpful starting place. Would definitely love to attend another discussion. Been reading all the books but hearing from others out there doing the work is in the government sphere is helpful too.

  • Yes to the point about efficiency! For public sector workers, AI has the potential to accelerate austerity. I love that the name of this session includes "the Myth of Inevitability"--nothing about it is inevitable. 

  • Thank you all for the discussion today! I've been struggling with variability in approaches/attitudes at different levels of my organization (in state gov, so there's statewide policy, agency guidance, and directions from team leadership) --- pushing back at the most immediate level doesn't always seem to be effective, even when framing it in terms of what's been published by higher levels of our org. It's hard to feel a sense of futility and I'm grateful for the solidarity from folks here today. 

  • Also, AI tends to make inherently biased decisions. AI also tends to change so rapidly that the amount of time to keep up with AI literacy and properly review its outputs for most work may actually make folks less efficient.

  • I feel like the point about conservative AI vs. innovation and creativity is super interesting - it makes me think of the question - why are we putting this emphasis on efficiency and speed at all? Is quicker always better? 

  • AI tends to make inherently biased decisions. AI also tends to change so rapidly that the amount of time to keep up with AI literacy and properly review its outputs for most work may actually make folks less efficient. 

    • ^my agency just instituted mandatory AI training by InnovateUS, which was funded by Google. It had some good information about specific ethical considerations, but it definitely had a pro-AI bias

      • ^the state of CA mandated one that I had to take seemed more like AI propaganda than an actual training. I was very disappointed in it

  • I think a major challenge is lack of fluency in how the tools work, I'm seeing staff get duped by vendors trying to push products to "improve" core services, and they don't know enough to ask tough questions 

  • I've asked every vendor I've seen demo a product what safeguards they have against discrimination or bias and not one of them has had a real answer 

    • ^because there aren't any. Most of the time, they don't even know what the model was trained on or they can tell you because they don't want to disclose that they stole everything

#GAREHosted
#EnvironmentalJustice
#DigitalEquity

#Public-Recordings

Statistics
0 Favorited
203 Views
2 Files
0 Shares
18 Downloads
Attachment(s)
pdf file
GARE Session Slides_8.6.26.pptx.pdf   4.22 MB   1 version
Uploaded - 08-27-2026

Related Entries and Links

No Related Resource entered.