What should the social contract be between the communications sector, the government, and the public in the next two decades?

Its terms should reflect that the purpose of communications policy is to ensure that all Americans can fully participate in the economy and civic life of their communities.[1]

But to understand why, let’s review our country’s approach to that question in the last three decades.

In 1996, there was a bipartisan consensus that policy should drive competition that, in turn, would drive the communications sector to continually deliver faster, better and cheaper communications for all.

Many thought it was mostly about voice—with incumbent local telephone companies (ILECs) competing against long-distance (IXCs) and new, competitive local phone companies for that market—or video, with the primary competition being between cable and satellite.

The then Federal Communications Commission Chairman Reed Hundt thought it was about digital; that is, digital would consume the siloed voice and video markets, leading to more competitive markets as voice and video became applications riding over digital transmission networks. By 2009, with a new Administration developing its framework for how to approach telecommunications policy, nearly everyone in the field understood that assessment to be right.

The FCC built the 2010 National Broadband Plan on that understanding, along with prior policy goals to propose a social contract under which policy would ensure broadband was everywhere, everyone could afford to be on it,[2] and we would use broadband to improve the delivery of all essential public services.

We could argue about the extent to which we succeeded,[3] particularly in getting competitive networks everywhere.[4]

But there should be no argument about this: in the wake of developments with AI, it is time to think again about the social contract.

What should that social contract be?

Let’s start by reviewing four lessons from 1996 and 2010.

1. Understand Evolving Incentives

One lesson is that with every technological transition, enterprises reorganize to build the capabilities they believe will put them in a position to earn a premium return on investment (ROI).

As the competitive landscape shifted in the wake of the ’96 Act and the digital transition, long-distance carriers, once they lost the ability to compete in the local phone service marketplace through what was called Unbundled Network Elements-Platform (UNE-P), essentially gave up and sold themselves to the large ILECs. The large ILECs also bought up many of the mobile carriers. If we could have kept them from doing so, the ILECs would have invested more and sooner in fiber, and wireless would have also invested faster to upgrade, enabling more intense competition with the ILECs and cable. But the product markets, when those mergers occurred, appeared separate, so the antitrust grounds for blocking such deals would have been weak. Moreover, there is no doubt that positive synergies exist with ILEC/IXC/mobile services.

But perhaps the biggest lesson is from the mergers that did not produce those synergies: specifically, ISPs and content. We can argue about why such combinations did not last long. Still, one cannot look at the record of US West/Time Warner, AT&T/Time Warner Entertainment, Verizon/Yahoo, Time-Warner/AOL, News Corp./DirecTV, and now Comcast/NBCU and argue that the market sees an economic upside of such combinations.

One conclusion we should reach is that the risk of distributors using their networks to discriminate against or suppress content creators is likely low to non-existent. That suggests the effort to mandate net neutrality—an effort that dominated public policy debates for two decades—is now obsolete. In an era of bandwidth scarcity, networks may have had incentives to use that scarcity to discriminate. When AT&T’s CEO threatened to charge the edge content providers for access, the average download speed was about 1.5 Mbps. Now, it is over 300 Mbps.[5] Consider the analogy of vehicle, not internet, traffic. We inside the beltway know the temptation to use the fast-pass lanes when facing bumper-to-bumper traffic in the four free lanes. But would any of us pay to do so if there were 400 free lanes?

Nope.

And that is the problem for those who thought that combining content and distribution would lead to a premium return.

The ROI for discrimination may have existed in the past, but it doesn't today.

In short, Comcast spinning off NBC/U should be seen as one in a long line of examples of how the content and broadband distribution industries are reorienting themselves to an age of streaming video and AI.

And just as Comcast has adjusted, so too should public policy advocates.[6]

In short, major technology and/or policy shifts can also push enterprises to adjust their assets, which affects how they serve their private goals as well as public policy goals.[7]

2. Understand the Advantages of Incumbency

Another lesson of the last three decades is that innovation requires us not to let those with huge historic advantages use them to win the transition.

In the 1990s, the ILECs held a monopoly over terminating access for voice services. The risk was that they would exploit that position to maximize profits while stifling mobile and data services—setting terminating access rates high enough to reduce the value and adoption of both.

We responded by creating obscure but critical policies such as lowering terminating access charges for data and mobile.

Another obscure bottleneck was who owned the phone numbers. That led to the adoption of wireless number portability, without which consumers switching carriers would have had to give up their phone numbers—friction that would have reduced competitive intensity.

It is not clear where such bottlenecks are in the AI ecosystem.[8] What is clear is that every major company in the space will seek to develop or buy such bottlenecks, and policy should monitor the landscape to prevent them from constraining innovation and growth.[9]

3. Use Public Investment Judiciously

A third lesson relates to public investment.

Most would agree that in the last three decades there was a need for public investment in at least five areas:

1. Public investment to support deployment in areas where the private sector is unable to achieve what Wall Street considers an acceptable ROI;
2. Subsidies to make broadband affordable for low-income persons, as otherwise, they cannot participate in the economic and civic life of their communities;
3. Federal subsidies for anchor institutions—schools, libraries, and health care facilities—that otherwise would not have sufficient information technologies to achieve their missions;
4. Support for training persons to use communications technologies to improve their lives; and
5. Government efforts to use the technologies to improve the delivery of essential public services.

Some go further and suggest that the government should invest more directly to compete with private-sector investments, particularly in broadband networks.[10]

Calls for government investment in government-owned broadband networks have been decreasing over the last few years. Still, as discussed below, there are calls to invest in government-owned AI-related facilities.

4. Be Wary of the Harms of New Technology

A fourth thing we learned is that we underestimated the potential harm of the new technology.

I could cite many examples, but the recent nearly $18 billion settlement between Meta and the states for harms caused to children illustrates many things.

For one, Meta stock rose as investors thought the settlement was both small and clever in how it might force competitors to adopt new practices.

From the public’s perspective, however, it was a confession that Meta had been poisoning our children.

Yes, as a legal matter, Meta did not admit to anything. But the public knows that no company pays $18 billion without being guilty of something.

It illustrates one approach: using tort litigation to address harm. That is better than nothing, but we could regulate to prevent the harm from ever occurring.[11]

We should remember that in 1990, the government acted to prevent certain kinds of harms, such as limiting advertising on television shows geared for children.

We did not do that with the internet, though California just passed a new regulatory framework to protect kids and others from potential AI harms.

We could argue about why that delay in reacting to demonstrated harms occurred. Still, the point is that we should consider protections for all, especially children, in a world dominated by AI.

Looking Forward

So, thinking about those four lessons, what should the framework for a new social contract be?

The framework should continue to ensure a market that produces faster, better, and cheaper communications for all.

It should build on the progress of networks everywhere with everyone on and using the networks to improve the delivery of essential public services.

But, to repeat what I wrote at the beginning, the social contract must reflect that the purpose of communications policy should be to ensure that all Americans can fully participate in the economy and civic life of their communities.

To develop a framework for that contract, the FCC, at a minimum, should undertake Notices of Inquiry (NOIs) on the following topics:[12]

• Information and institutions—What information is needed to continually improve policies about the opportunities and risks of the AI economy and society? What new powers does the FCC need, or should there be a new expert agency, to address the issues we know we will face in the AI economy and society?
• Networks and other facilities—What communications networks do we need for our country to thrive in the AI economy? What are the gaps in making such networks accessible and affordable to all? How should we reform the Universal Service Fund—as we did when we moved from voice to mobile and then to broadband—for an AI economy and society?[13] What are the investments the public should be making to assure public benefits and protection in the AI economy?
• Protections—How do we protect children and others from potential AI harms? How do we assist consumers to make informed decisions about which AI products they use that can propel a race to the top rather than a race to the bottom? How do we protect against cyber and national security risks?
• Government Performance—How can the tools that underlay the AI economy improve the delivery of essential public services, including education, health care, and public safety?

Then, to the extent that the debate supports actions the FCC can take, it should do so. If government action is needed but the FCC lacks authority, the FCC should send Congress a to-do list.

You might ask what the answers to these questions are?

I could tell you, but I am approaching my word limit.

I’m kidding.

I don’t know the answer to any of the questions.

When I started working on the National Broadband Plan, I didn’t know the answer, except for what my life has taught me: getting the optimal answer requires asking the right questions to a broad group of stakeholders and then analyzing their responses.

But let me offer a couple of thoughts on the question of public investment, as there has recently been increased interest in whether public funds should be invested in data centers.[14]

I understand there to be three principal arguments:

1. That data centers will end up like utilities with monopoly powers and public ownership is preferred to utility regulation.
2. That data centers use public resources—including our collective knowledge—and therefore, the public should have greater control and ownership.
3. That data centers should give smaller players—researchers, startups, and civil society—greater access.

Wall Street and the tech community would no doubt immediately oppose such actions, arguing that public investment should not compete with private investment.

I don’t have the same ideological reaction, and I think the proposals are worth studying.

Still, my question is: What does public investment get us that private investment does not? There are some possibilities. For example, public facilities might be more sensitive to privacy, environmental, or other concerns that help prevent collateral harms. But I think government regulation is better suited to address those concerns and prevent such harms across all data centers. I think the concerns about bottlenecks and monopoly power are worth monitoring, but the investments we are seeing reflect significant competition.

As to what we might view as public access—access to use data centers in ways that produce a public good but not an acceptable ROI—I think multiple paths short of public ownership can achieve that result.

But I don’t really have an answer. The debate is timely but still underdeveloped.

I believe, however, that we are quickly approaching the time when Congress will start drafting the telecommunications/social media/internet/AI regulatory reform act.

2029 is more likely than 2027 for such an act to pass.

Now, however, is the time for the work that this conference, A New Compact for Connectivity, is doing: asking the right questions that allow us to develop defining policies that improve the lives of all Americans in the face of the opportunities and threats that the AI economy presents.

Blair Levin is the Policy Advisor to New Street Research and a nonresident senior fellow at Brookings Metro. Prior to joining New Street, Blair served as Chief of Staff to FCC Chairman Reed Hundt (1993-1997), directed the writing of the United States National Broadband Plan (2009-2010), and was a policy analyst for the equity research teams at Legg Mason and Stifel Nicolaus. Levin is a graduate of Yale College and Yale Law School.

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