You’re reading the Benton Institute for Broadband & Society’s Weekly Digest, a recap of the biggest (or most overlooked) broadband stories of the week. The digest is delivered via e-mail each Friday.

Round-Up for the Week of June 8-12, 2026

This week, the Brookings Center on Regulation and Markets held a fireside chat with Neil Thompson, director of the FutureTech project at the Massachusetts Institute of Technology (MIT), to explore the intersection of artificial intelligence (AI) and economic mobility. The conversation between Thompson and Sanjay Patnaik, Director of the Center on Regulation and Markets at Brookings, is part of an ongoing Brookings series, the Economics and Regulation of Artificial Intelligence and Emerging Technologies. The fireside chat highlighted a number of pressing issues related to the development of AI technologies, including the progress of generative AI models, workforce impacts, potential regulatory avenues, and whether AI at large will lessen or deepen current inequalities.

AI is Impressive, but is it Reliable or Affordable?

The goal of FutureTech is to understand what's driving AI models' development and what effect they might have on the labor market, the size of the economy, and science. Part of this is separating the hype around AI models—such as OpenAI's ChatGPT or Anthropic's Claude—from the real-world evidence that shows where AI sits in the technology adoption curve. While many AI models are advancing rapidly and improving in quality, Thompson warns that the difference between "pretty good" and "perfect" performance is critical for the impact of widespread AI model use.

"People have done work where they say, 'Oh, well, let's use generative AI to summarize documents.' And you find that sort of three to seven percent of the time, [the models are] just adding new information that was not in the underlying document in the summary," said Thompson. "It's easy to look at a summary and say, 'Yeah, that looks about right.' That is not the same as it is exactly right, and that's going to matter a lot as we start to think about how [AI models] get in the economy."

Thompson compared AI's rapidly increasing prevalence to the technological revolution from a Model T to a modern-day Ferrari. To Thompson, AI models are being scaled at an exponential rate, and the costs of that development increase exponentially in tandem. That pace may not be sustainable, and tools that improve due to more funding and compute—the specialized hardware and infrastructure required to train, refine, and run AI models—being thrown at them may become too expensive, too fast.

"About eighty percent of the improvement that we see in [AI models] just comes from throwing more computing power at them," he said. "That means that we are in that Model T, going to the Ferrari world. It's absolutely impressive what they're doing, but we can't always just keep scaling up the cost of these things in order to do more with them."

AI's Impact on Innovation

Thompson sees innovation around AI models as critical to supporting innovation more broadly. With the influx of investment comes an expansion of innovative outcomes that, ideally, boost prosperity. While AI developers say the technology will be adopted broadly and quickly, diffusing across the economy, public opinion has increasingly turned negative toward AI models. But will it all happen on the schedule AI developers propose? Thompson says, maybe not.

"There [is] this thought of, 'We're going to adopt these. These systems are going to be great.' And [there are] these predictions about big productivity gains," Thompson said. "And actually, if you look at the empirical evidence, it's pretty weak about the size of the productivity gains that people are getting. But it's not because it's not changing things. It's just that there's positive and negative happening at the same time."

Productivity gains because of AI models will come with further development, Thompson added. Because these systems are so new, some time is needed to iron out the kinks. How to make these systems better, he said, is a critical question to be answered.

"How do we make sure that we keep getting more of the benefits and we develop systems that help us reduce the downside?" Thompson said.

"We need that framework around it," Patnaik said, "that institutional and policy framework to make sure that we can actually deploy these technologies to the, to the best benefit. "

AI, Learning, and Labor

AI models are in classrooms across the country and across all ages. As a result, the learning environment is changing drastically. Patnaik and Thompson noted that the question of how best to implement AI tools in learning environments is very tricky, and that both students and teachers are wondering how to move forward as AI tools develop.

Patnaik said that the foundational communication skills humans possess will remain as important as ever. Currently, people are facing the challenge of AI literacy as an extension of basic literacy, something that may be masked by the technology.

"Whether you communicate with an AI or with a human, you need to be able to express yourself … in written form," he said.

When asked how AI connects to preexisting digital divides, Thompson said AI’s impact is a complex challenge for which he does not yet have an answer. However, he added that AI diffusion will add more complexity to how the digital divide is currently understood.

"You might have kind of like a blurring of the lines in terms of how different communities get affected," responded Patnaik.

Regarding the future workforce, Thompson said the challenge will be redistributing labor as a result of the changes AI models will make to how work is currently done. This challenge could have international implications, he said. In countries whose economies are built around narrow industries that have been able to lift people out of poverty, they could lose those industries if AI takes them over.

"AI is going to automate a bunch of things, right? For a big economy like the United States, that means there are going to be winners and losers, but hopefully, we can build systems within the economy that will redistribute people and figure out how they can do other jobs," Thompson said. "That's much, much harder if you have a giant concentration of people all doing the same thing. If AI hits that, there are gonna be really profound disruptions, and that's one I worry about."

Moving Forward With Regulation

Patnaik asked Thompson how states should approach regulating AI technologies amid tension between existing state legislation and the federal government's deregulatory stance. Thompson said that a recent FutureTech survey of AI experts found they were concerned about substantial economic and health risks arising from widespread AI use.

"That level of risk calls for people creating standards, creating regulations, to make sure that some of the worst outcomes that can come from [AI development] are going to be mitigated as they go," he said.

Thompson also added that antitrust enforcement will be critical in ensuring fair competition going forward.

To this, Patnaik asked how regulators can keep pace with newer and newer AI models and capabilities, responding as close to real time as possible. According to Thompson, the question is more about understanding the deeper trends at play and taking a long-term view.

"What we want is to be able to create laws and regulations that are not for somebody's anecdote about what AI was like two years ago," Thompson said. "We want to know the trajectory that these systems are on, and therefore, [what] is going to change. Let's prepare for those things today. I think that is possible, but it's complicated because [legislators] have to understand some of the technical pieces going on." 

Quick Bits

• States Must Consider Future of Broadband Offices
• Three more providers default on RDOF awards
• The Story of U.S. Spectrum Policy
• White House, Hill relaunch effort to block state AI laws

Weekend Reads

• Artificial Intelligence in the National Security Enterprise
• Broadband in America Report: Cable Market Focus

ICYMI from Benton

• The FCC Is Asking Whether to End E-Rate. Every School and Library Should Answer.
• Digital Access Is About People: My Experience as an American Connection Corps Digital Navigator
• Marking the Paper Anniversary of the BEAD Restructuring Notice
• President Trump Signs Directive Reshaping How the Military and Intelligence Community Use AI

Upcoming Events

June 12—AI Cybersecurity After the Executive Order: What's Next? (Congressional Internet Caucus)

June 16—Policy Outlook on Data Privacy & Cybersecurity (Multicultural Media Telecom and Internet Council)

June 16—How to Protect Kids From Chatbots Without Bans (Information Technology & Innovation Foundation)

June 17—Community Connect (Department of Agriculture)

June 24—Building for AI: The Next Wave of Infrastructure Investment (Georgetown University)

June 25—June 2026 Open Federal Communications Commission Meeting (Federal Communications Commission)

June 30—The New Push for a National Data Privacy Standard (Information Technology & Innovation Foundation)

The Benton Institute for Broadband & Society is a non-profit organization dedicated to ensuring that all people in the U.S. have access to competitive, High-Performance Broadband regardless of where they live or who they are. We believe communication policy - rooted in the values of access, equity, and diversity - has the power to deliver new opportunities and strengthen communities.

© Benton Institute for Broadband & Society 2026. Redistribution of this email publication - both internally and externally - is encouraged if it includes this copyright statement.

For subscribe/unsubscribe info, please email headlinesATbentonDOTorg

Kevin Taglang
Executive Editor, Communications-related Headlines
Benton Institute
for Broadband & Society
1041 Ridge Rd, Unit 214
Wilmette, IL 60091
847-220-4531
headlines AT benton DOT org

Share this edition:

Broadband Delivers Opportunities and Strengthens Communities