The divide, by the numbers

• Internet use in the U.S. rose to83% of people ages 3+ in 2023, up from 80% in 2021, with notable gains among historically underserved groups. That’s progress—but it still leaves tens of millions outside full participation. (ntia.gov)
• Gaps at home remain stark: American Indian and Alaska Native households have the lowest home broadband rates (75%), with Black and Hispanic households also lagging White households. (federalreserve.gov)
• Policy headwinds matter. The FCC’s recent vote to end discounts for library hotspot lending and school-bus Wi-Fi programs removes lifelines many low-income and rural families relied on to get online for school, work, and services—changes advocates warn will deepen the divide. (apnews.com)

Why AI makes the divide more expensive

Early, high-quality evidence shows generative AI materially raises worker output:

• A large-scale deployment of a generative-AI assistant to 5,179 customer-support agents increased issues resolved per hour by 14%, with novice/low-skill workers improving by 34%—a strong equalizing effect for entry-level talent. (nber.org)
• Controlled experiments on professional writing tasks find faster completion and higher-quality outputs using ChatGPT-style tools. (economics.mit.edu)
• The OECD’s 2024 synthesis concludes AI has credible potential to revive sluggish productivity growth—if micro-level gains scale across firms and sectors. (oecd.org)

In other words, AI tools are force multipliers. Workers who have access to connectivity, devices, and basic AI fluency will produce more, faster. Those without will fall behind faster than in any prior tech cycle — because this cycle directly augments cognitive work across roles.

The GDP link: connectivity → skills → productivity → growth

The relationship between connectivity and economic output is well-documented:

• Studies across countries show that a10-point increase in broadband penetration is associated with roughly a 1% increase in GDP(directionally, magnitude varies by context). (adlittle.com)

AI is amplifying that logic. Connectivity is now the on-ramp not just to information, but to AI co-workers—coding copilots, writing assistants, analytics copilots—whose benefits compound with use. As the AI build-out accelerates (e.g., data-center investment), analysts are already attributing measurable contributions to GDP growth from AI-related capex, with expectations of further tailwinds in 2025–2026. (reuters.com)

Takeaway: If households and small employers can’t get—and keep—robust broadband, they can’t plug into the productivity gains driving growth. The macro cost is foregone output; the micro cost is stalled wages and reduced employability.

What this means for the future workforce

1. Entry-level acceleration—or exclusion. Evidence shows the biggest AI-driven productivity gains accrue to newer, less-skilled workers—exactly the populations we need to onboard quickly. Without access, these workers lose the very “lift” that could close skill gaps. (nber.org)
2. Skills become “connected skills. AI literacy depends on reliable internet, modern devices, and safe access. Without that triangle, classrooms, training programs, and employers can’t integrate AI into day-to-day learning and work. (ntia.gov)
3. Regional competitiveness at risk. Communities with uneven broadband will see fewer firms adopting AI workflows (or doing so unevenly), widening productivity and wage dispersion between connected metros and left-behind counties. (oecd.org)

The cost of inaction

• Under-investment in last-mile access now implies persistent headwinds to productivity growth just as a new GPT (general-purpose technology) arrives. OECD’s modeling underscores the stakes: the macro payoff requires broad diffusion, not isolated pilots. (oecd.org)
• Program retrenchment(e.g., ending support for hotspots/school-bus Wi-Fi) removes practical bridges that kept students and families online, especially where fixed broadband is unaffordable or unavailable. Expect downstream effects in homework completion, job applications, telehealth, and skills attainment. (apnews.com)
• GDP leakage. Historic estimates linking broadband penetration to GDP growth suggest that failing to close adoption gaps systematically caps output—and now, AI compounds that cap by making connectivity a prerequisite to productivity tools. (adlittle.com)

What to do next (policy, philanthropy, employers)

1) Protect and scale practical access bridges. Libraries, schools, and community organizations are critical distribution points for connectivity and devices (hotspot lending, community Wi-Fi, device refurb/reuse). When these supports are pulled back, local leaders and private partners should step in with public-private funding to sustain them. (apnews.com)

2) Fund “AI-ready” broadband and devices, not just minimal access. The bar has moved from “any internet” to reliable, low-latency service that supports real-time AI tools, uploads, and collaboration—plus modern devices for households and frontline workers. (NTIA’s data explorer can pinpoint local usage/adoption gaps to target.) (ntia.gov)

3) Make AI literacy the new digital literacy. Layer AI-basics (prompting, verification, privacy) into workforce programs, high schools, and community colleges. The goal is to convert access into measured productivity gains—exactly the improvements validated in field and lab studies. (nber.org)

4) Employers: deploy copilots where they help your newest people. Prioritize AI rollouts in functions where novice employees benefit most (support, sales, ops documentation). Track issues resolved per hour, time-to-proficiency, and quality scores to quantify ROI and close skills gaps faster. (nber.org)

5) Leverage Premier’s Hotspot Checkout Program and Premier’s CPR3 program as strategic tools for adoption. Premier Wireless offers two proven ways for organizations to bridge the connectivity and adoption gap:

• Premier’s Hotspot Checkout Program enables schools, libraries, nonprofits, and government agencies to loan Wi-Fi hotspots directly to community members—putting high-speed internet access in their hands immediately, bypassing the delays of infrastructure expansion.
• Premier’s CPR3 program, powered by T-Mobile—America’s largest and fastest 5G network, is an exclusive Premier Wireless solution that Connects People to Resources. CPR3 effectively closes the digital divide by providing high-speed internet access, a connected device, and curated access to education, skill training (including AI adoption), and educational-grade content filtering.

Together, these initiatives represent a scalable bridge from no connection → basic access → digital enablement and AI engagement. They allow communities to go beyond connectivity to true participation in the modern, AI-enabled economy.

Bottom line

We’re at the opening chapter of a new industrial revolution. The early evidence is clear: AI increases output—especially for the people who need the biggest boost. But that promise only reaches them if they can get online, stay online, and learn to work with AI. Closing the digital divide isn’t just a social good; it’s a growth strategy for America’s workforce and GDP.