Before We Talk About AI in Classrooms, We Need to Talk About Who’s Still Offline

---

While headlines debate the promise and peril of AI-powered classrooms, millions of students still don’t have reliable internet at home. A 2018 Pew survey found that 24% of teens from families earning under $30,000 lacked dependable internet for homework, compared with 9% among higher-income peers, and that one in four Black teens were sometimes unable to complete assignments because of it, compared to just 4% of white teens. The digital divide was a well-documented problem long before the pandemic made it impossible to ignore. And despite years of federal investment aimed at closing it, progress has been uneven and fragile.

Infrastructure and funding are part of the answer, but they aren’t the whole story. A significant yet frequently overlooked obstacle is trust. In its absence, even the most well-funded digital equity initiatives are likely to be ineffective.

The Gap That Keeps Coming Back

Federal programs have made real progress in expanding access. Initiatives like E-Rate, ConnectED, and the Broadband Equity, Access, and Deployment Program have brought high-speed internet to schools and communities once considered too remote or too costly to serve. But that progress has never been steady.  Funding lapses, shifting administration priorities, and high service costs have repeatedly stalled implementation and left many schools and families without reliable access.

During the pandemic, when connectivity workarounds like completing homework on a phone screen or driving to a library parking lot for Wi-Fi became unsustainable, the federal government launched the Emergency Connectivity Fund and the Affordable Connectivity Program to help cover the cost of devices and home internet for eligible households. These programs worked for awhile, but when funding expired, many families lost access again, and the achievement gaps that had briefly narrowed began to widen once more. Research from Michigan State University’s Quello Center found that students in rural Michigan with broadband access had, on average, a 0.6-point higher GPA than those without.

Access alone, however, only goes so far. Even where infrastructure and subsidies exist, many households lack the training to put digital tools to use. Research shows that a student with digital skills even modestly below average is 29% less likely to plan to complete a college or university program. The consequences extend beyond school. According to the National Skills Coalition, 92% of U.S. jobs now require digital skills, but one-third of workers lack them, a gap that begins forming in early education. The Digital Equity Act of 2021 created grants to address this by funding state-level digital literacy and skills programs. However, when the Trump administration ended the program in 2025, it became yet another example of an initiative that was discontinued before communities had a chance to reap its full benefits.

Why People Don’t Trust the Programs Meant to Help Them

That pattern of interrupted investment has consequences beyond the immediate loss of services. It erodes trust, and the communities most affected by the digital divide are already among the least likely to trust the institutions promising solutions. A 2023 Gallup poll found that only 26% of Americans expressed confidence in major institutions. The 2024 Edelman Trust Barometer found that trust is even lower among low-income Americans. Only 32% trust government and 38% trust business, compared to over 50% among higher-income groups. And research from Data & Society found that 48% of online adults without a high school degree say they trust their internet service provider “only a little” or “not at all,” compared with 33% of college graduates.

This distrust isn’t irrational. It reflects lived experience with programs that promised lasting change and delivered temporary relief. And it creates a real obstacle for digital inclusion efforts, because people who don’t trust the institutions behind a program are unlikely to participate in it, no matter how well it’s funded.

What tends to work instead is more local and more personal. Librarians, teachers, and volunteers at churches and community centers often serve as the actual bridge between residents and digital resources, offering guidance that builds genuine confidence rather than institutional skepticism. A 2022 Boston Consulting Group report found that 85% of people who received help from digital navigators, local on-the-ground supporters, said they now use the internet more frequently. Trust, in other words, is built through relationships, consistency, and credibility at the community level, not through top-down program rollouts.

Measuring What Actually Matters

The success of community-based models points to a gap in how digital inclusion is currently evaluated. Most policies measure progress through technical indicators: broadband speed thresholds, device counts, household enrollment figures. These metrics matter, but they capture access, not participation. A household can be technically connected and still effectively excluded if the people in it don’t have the skills, the confidence, or the trust to use digital tools in meaningful ways.

Treating trust as a measurable outcome, alongside connectivity, would give policymakers a more complete picture of where digital inclusion efforts are actually working. Community trust surveys could help identify who builds confidence in digital systems and how, whether through personal relationships, cultural alignment, consistent follow-through, or transparency. That kind of data could then inform more targeted investments in both infrastructure and the community anchors who do the harder work of making that infrastructure feel accessible and worth using.

As AI becomes more integrated into classrooms and public systems, the stakes of getting this right will only grow. The same communities that have been underserved by the digital economy are the ones most at risk of being left behind by the next wave of technological change. Trust won’t solve the digital divide on its own, but without it, even the best-funded programs will keep running into the same walls.