Minnesota forum reveals widespread unease with AI technology race

This article was originally published by Minnesota Reformer.

From job loss to nudification technology, data centers to the end times, a recent southwest metro forum on artificial intelligence underscored the low-boil anxiety of Minnesotans about this unsettling technological change — and myriad policy questions emerging from it.  

A recent campaign forum-style event at New Prairie High School in Eden Prairie was moderated by Dr. Krish Subrahmanian, a Hennepin Healthcare pediatrician who is also the Democratic-Farmer-Labor nominee to succeed state Sen. Steve Cwodzinski in a district covering parts of the southwest Twin Cities metro. He was joined by Attorney General Keith Ellison, as well as representatives from academia, business and labor. 

Most of them brought up the unsettling pace of advancement at the AI frontier. After a researcher with Claude developer Anthropic publicly resigned earlier this month with a warning that humans could soon lose control of self-improving AI if development isn’t throttled, top executives from Anthropic and OpenAI called for a mutually agreed slowdown — though the industry remains split on the issue, and lawmakers are unlikely to take it up before the Nov. 3 election with the House’s hasty exit from Washington last week.

Kevin Pranis, business and marketing manager for the LIUNA trade union in Minnesota and North Dakota, said he hoped the stakes would help bring people together. 

“We should be taking (it) seriously as a society, and really start focusing and maybe come together around something. … This is probably not a good one for partisan battles,” he said.

It’s encouraging that insiders are raising concerns externally because it shows that the industry is not united in pushing the frontier at all costs, said Obasesam Okoi, director of the AI and Society Lab at the University of St. Thomas. 

“To me it means we are getting somewhere. Now the change is emerging from within organizations and companies that are designing these technologies that humanity is still trying to figure out where exactly is this race taking us,” Okoi said.

Speakers urged attendees — and policymakers — to focus not only on the threat of a catastrophic loss of human control but on less dramatic AI risks, too. 

Okoi said an awareness of “unintended consequences” should be baked into the development of AI models used in areas where improper use can cause harm, like policing and medicine. In November, the Brennan Center for justice warned that police departments’ unrestricted use of predictive AI could increase error rates in policing decisions while embedding social biases. AI diagnostic tools, meanwhile, are at the heart of the recent doctors’ strike at Allina Health System.

Humans are ultimately responsible for how AI models are used in the real world, Ellison said, calling out Elon Musk of SpaceXAI and OpenAI CEO Sam Altman by name. Earlier this year, OpenAI admitted one of its models broke out of a training “sandbox” and hacked another AI company; its tools have since been implicated in other breaches.

“No, AI is not taking over. Maybe Sam Altman has taken over. Maybe Elon Musk has taken over. Human beings make mistakes, and we should be able to challenge them,” Ellison said.

The state consumer data privacy law that took effect last year — and which Ellison’s office enforces — gives Minnesotans important new rights that can help protect against AI abuses, such as biased decisions by automated tools that make medical, housing and employment decisions, he added.

Nonetheless, more regulation will likely be needed, Okoi said.

“If risks are high, then regulations have to be high,” he said.

Even existing regulations are being challenged, however, as in the case of Musk’s company suing Minnesota over a law intended to prevent the creation of sexually explicit images or videos of a person without their consent.  Ellison pledged that he will successfully defend the state’s first-of-its-kind anti-nudification law against a legal challenge from Musk’s AI company, which argues on free speech grounds.

Ellison’s attorneys notched two key procedural victories in the case earlier this summer, and the matter is heading to trial, he said at the forum. 

“I have every reason to believe we’re going to defeat it again,” Ellison said.

Subrahmanian, who is faces Independent Kris Babler and Republican Dustin DuFault in the upcoming Senate District 49 election, also asked attendees to weigh in on the physical avatars of AI: data centers. 

Astrid Mozes, chair of the Swedish American Chamber of Commerce in Minnesota, said data centers could partially assuage concerns about energy use and environmental impacts — including the massive amounts of heat shed by their cooling systems — by hooking into district energy networks that provide domestic heat and hot water for millions of homes and businesses in Europe and in some U.S. cities, including Minneapolis and St. Paul.

Executives from Ever-Green Energy said earlier this year that it would be feasible to reuse data centers’ waste heat in systems like the one it operates in downtown St. Paul.

“(The) technology is available, solutions have been implemented. We just need to replicate it,” Mozes said.