Jessica Dine: We Have the Tools to Head Off the Newest Digital Divide
If the NTIA would let us use them...
In recent news, Starlink—whose satellite broadband service won bids to cover 470,000 locations subsidized by the federal government’s BEAD program—is now asking states to exempt their networks from meeting basic requirements like ongoing performance testing in certain served locations, reservation of network capacity, and award payment schedules that are tied to subscriber milestones.
While the National Telecommunications and Information Administration boasts about having cut $20 billion from the Broadband Equity, Access and Deployment program through its dubiously-legal restructuring, the state officials tasked with implementing BEAD are scratching their heads, trying to figure out how to move forward with ISPs that want a guaranteed payout—whether people get online or not.
It’s not only internet access at stake anymore. Even as we struggle to close the original digital divide, a new, immensely consequential one has opened up beneath our feet—the gap created by uneven access to and use of AI.
As generative AI becomes embedded in everything from educational software and hiring systems to loan applications, Americans are starting to need both baseline digital skills and basic AI proficiency to navigate daily life. And the same populations already struggling to get online are now being left behind by the AI revolution. The window to address these gaps is closing fast.
Good thing Congress already addressed much of this in the original BEAD rules. If only NTIA were willing to employ them.
One of the most contentious parts of BEAD has been whether states have the discretion to spend funds on “non-deployment” uses—things like subscription affordability programs, initiatives to increase broadband adoption rates, device acquisition, and digital upskilling such as cybersecurity and privacy training. Originally, Congress granted states broad permission to use the remaining funds to fill any gaps that are left over after BEAD deployment takes place.
Notably, NTIA temporarily revoked that permission shortly after canceling the Digital Equity Act programs, which Congress intended to supplement BEAD with funding for digital adoption and readiness initiatives. And whether states will ever have permission to spend their remaining BEAD funds—or on what—remains to be seen. Although a decision is expected in early 2026, with an NTIA listening session being held in a few weeks, for now, NTIA’s intentions are opaque and mixed at best.
Today, the need for those funds is clearer than ever. When Starlink—which won bids to serve nearly 12 percent of BEAD locations (22 percent combined with Amazon’s Leo)—pushes to be reimbursed based on self-certified “service availability” rather than actual consumer take-up rate, it’s ignoring the fact that its $80 monthly fee is too high for many consumers and that satellite adoption has historically been low.
That means states need the tools to address those affordability and digital skills gaps so consumers can get online. If Starlink won’t provide professional installation for consumer equipment that often needs to be mounted on roofs or other inaccessible places—and it can’t guarantee performance if the consumer installs it wrong—then states need the funds to offer that service themselves.
If internet providers won’t ensure consumers can actually benefit from the networks that are built, Congress intended for states to address those gaps themselves with the funding that NTIA has yet to release.
And the real kicker is that while NTIA waffles and state efforts to close the digital divide are stymied, the divide created by uneven access to AI keeps growing.
Soon, just as you need the internet to access education, work, financial services, and healthcare, you’ll need a degree of AI proficiency to navigate our world. And just as Congress allocated funds to address the gaps and barriers to digital use, those same funds could be used to train and prepare populations to use AI as it becomes more relevant to daily life.
The same network of digital navigators, digital inclusion organizations, and other social hubs like libraries and schools could be used to support the population’s use of AI if they were given the resources to expand and update their curricula.
Bringing the offline population online isn’t just about laying fiber in the ground or beaming data down from satellites. The benefits of connectivity only accrue when people actually use those networks. That means addressing affordability gaps, reservations and uncertainty around using the internet, and other barriers that keep Americans offline. And a big part of the problem has long been gaps in knowledge and digital skill levels—some as basic as turning on a computer or sending an email—that make people wary or incapable of using the internet for what they need to do.
AI proficiency requires those baseline skills, along with a host of new ones: Things like how to differentiate between real and artificially generated images, how to use your data wisely, how to source information in a world where it can be hard to distinguish between AI hallucinations and fact. Even if you aren’t seeking out AI, you’re unlikely to avoid dealing with it altogether.
The same people who routinely run into problems with basic digital tasks are already disadvantaged when it comes to using AI. And the problem is compounding. While well-resourced schools and institutions provide training on using large language models, understanding algorithmic bias, and evaluating AI-generated content, the offline population is falling further and further behind. They’re being excluded from building generational knowledge, accrued familiarity, and accumulated comfort with new technologies that’ll be difficult—even impossible—to catch up on.
Compared to expensive broadband deployments, the expense associated with training populations to use digital technologies, including AI, is minimal. The costs of not doing so are enormous. For an administration focused on cost-cutting, the immense return on investment should be appealing. And the need to train Americans to make safe and effective use of AI should be compelling in a world where technological competitiveness is critical and fierce.
Congress already passed the law that made federal dollars available for states to spend on exactly this goal. It’s called BEAD. All NTIA needs to do is follow it.
Jessica Dine is a policy analyst at New America’s Open Technology Institute and Wireless Future Project, where she focuses on a range of issues including broadband access and adoption, spectrum policy, and competition. This Expert Opinion is exclusive to Broadband Breakfast.
Broadband Breakfast accepts commentary from informed observers of the broadband scene. Please send pieces to commentary@breakfast.media. The views reflected in Expert Opinion pieces do not necessarily reflect the views of Broadband Breakfast and Breakfast Media LLC.
Member discussion