Report warns AI could ‘entrench the status quo’ unless changes are made

Governments are moving quickly to embrace artificial intelligence, but a recent report urged them to rethink their entire structures and institutions, or else they risk repeating the mistakes of the past.

Researchers at RethinkAI, an initiative from the nonprofit New America, found earlier this month that there is massive interest in AI, especially among state and city governments, but its use remains mostly focused on optimizing existing processes and making them more efficient.

Instead, the authors called on public-sector leaders to pursue a different approach they referred to as ALT: Adapt, Listen and Trust, which would involve fundamentally reshaping their governments to be more responsive, adaptable and accountable in ways that restore and build public trust.

“The hard truth is that AI is already shaping government, mostly to optimize old processes,” the report says. “If we keep chasing efficiency, we’ll entrench the status quo.”

Neil Kleiman, a senior fellow and professor at the Burnes Center for Social Change at Northeastern University, said in a recent interview with Route Fifty that there has been a “real flip” in terms of policymaking for AI, as states have largely led the way on implementing and planning for the technology while many cities have lagged. That’s a sea change compared to previous tech trends, where cities took a leadership role, but state regulation has been primarily focused on mitigating harms and stepping up where Congress isn’t.

“The federal government has really ceded the territory of regulating this dangerous technology in a lot of people's eyes, so states feel like they have to fill the vacuum,” Kleiman said. “That's why you're seeing this tsunami of legislation over the last few years, but particularly the last few months.”

The report says much of that work in state legislatures has been defensive, “split roughly evenly between constraining government use and regulating private providers.” They also are investing in safe and secure environments where they can experiment with AI tools without risk. Cities, meanwhile, have focused more on pilot projects for specific use cases that address some of their most urgent problems.

New America identified six areas where city governments are finding AI use cases: permitting and automation, employee productivity, public safety, resident services, public engagement and infrastructure analytics. Researchers said that shows the “disparate range of challenges” facing municipal governments.

But they also warned it could be another false dawn, if AI is not handled right. Kleiman drew a comparison to the “smart city” movement of a few years ago, where municipalities invested billions of dollars into technology designed to make their cities more livable by dealing with traffic congestion, affordability and connectivity, among others, but failed in some people’s eyes to live up to that initial promise.

“In that rush for better, faster, cheaper, we didn't really focus on what residents really need, and we fell into an efficiency trap,” Kleiman said. “There's this centrifugal force to always be improving existing services rather than transforming those services, or rather than pausing and listening to what it is that citizens really need. In that regard, I think we kept moving in the direction of faster bureaucracy, rather than a better functioning government residents can trust.”

To avoid the mistakes of the past, the report recommended the ALT framework. The method would focus on how decisions are made by governments, not just what decisions are made, and New America argued it could help bring about what the report called “institutional reform.”

Researchers said adapting would mean anticipating future needs and repositioning institutions to meet them, while listening would mean committing to understanding what residents need. The trust aspect would be by governments holding themselves to account to their residents by showing results. Using AI to be purely for efficiency will just make for “faster bureaucracy,” New America said.

“The whole idea of this ALT framework is to try and compensate for ways that we didn't quite meet the aspiration of past tech waves,” Kleiman said. “In the past, we had what I call marginally better with technology solutions, and we think we can get dramatically better if we use AI the right way. If anything, I think in a lot of ways, AI could help local governments flip the fears on its head.”

Those fears include the public-sector workforce being hollowed out and AI-fueled misinformation being allowed to spread. Kleiman said he has seen a lot of interest among government leaders at all levels in doing things differently as they embrace AI. He said that gives him more optimism than with previous waves of technology.

“I have never in 30 years been in a place where I've seen local government more ready to fundamentally change how it engages with its residents,” Kleiman said. “I think it's really awakening, just in the last few months, to how AI could be a really powerful tool to do that.”