FCC Chair Faces Fine for Using Known Faulty ISP Data

FCC Chairman Ajit Pai cited BarrierFree’s data to highlight improved rural broadband access, even though the company’s figures had been flagged as potentially inflated.

The FCC launched the $20.4 billion Rural Digital Opportunity Fund last year as part of an effort to expand broadband to rural Americans who have long been underserved.

On its website, the FCC summarizes the scope of the challenge:

“In urban areas, 97% of Americans have access to high-speed fixed service. In rural areas, that number falls to 65%. And, on Tribal lands, barely 60% have access. All told, nearly 30 million Americans cannot reap the benefits of the digital age.”

Pai cited BarrierFree’s numbers in February of last year despite warnings that the provider’s deployment reports had gaps going back to 2015. He later acknowledged the mistake but continued to assert that broadband deployment was progressing in a “reasonable and timely fashion.”

At one point Pai suggested that the number of Americans without broadband fell from 26.1 million in late 2016 to 19.4 million a year later. Independent checks showed the true figure was closer to 21.3 million—highlighting a meaningful discrepancy.

The FCC has faced repeated criticism over the data it uses to measure progress in closing the rural connectivity gap.

Last year, Microsoft published a detailed critique calling the FCC’s broadband data misleading. Microsoft used telemetry from its own online services to estimate that roughly 163 million people “do not use the internet at broadband speeds.”

That statistic does not necessarily mean all of those people lack a broadband option, but the large divergence between Microsoft’s usage-based findings and the FCC’s coverage reports prompted further analysis.

Microsoft examined Washington state as a case study: FCC data suggested 100 percent of Ferry County residents had access to broadband. Local officials, however, told Microsoft that very few residents actually had access and that most broadband connections in the county were business-grade services. Microsoft’s usage data supported that local account, showing only about two percent of Ferry County residents were using broadband.

John Kahan, Microsoft’s Chief Analytics Officer, said similar discrepancies appeared across nearly every county in all 50 states, indicating systemic problems with the accuracy of access data reported to the FCC.

A core issue, Kahan noted, is the FCC’s reliance on self-reported information from internet service providers that often isn’t independently verified. When providers overreport coverage or subscriptions, the official data can make it appear the FCC is closing the digital divide more quickly than is actually happening—leaving residents of underserved communities with inadequate connectivity that limits economic, educational, and healthcare opportunities.

Democratic FCC Commissioner Jessica Rosenworcel warned of the consequences in a formal statement:

“It is no secret that the Federal Communications Commission has a big data problem. It may be the nation’s communications expert, but it doesn’t know with precision where broadband is and is not across the country. This is unfortunate, especially now when having access to high-speed service during a public health crisis is so critical.

During this pandemic, broadband is required for much of day-to-day work, education, healthcare, and more. But for too long the FCC has fumbled efforts to fix its broadband data and put off initiatives to improve its maps, making it more difficult to close the digital divide—both during this pandemic and beyond.”

This week the FCC proposed fining BarrierFree $163,912 for reportedly submitting inaccurate subscription counts that inflated its broadband numbers, failing to file required deployment data, making false statements to Commission investigators, and not responding to other inquiries.

The proposed enforcement action shows the FCC is beginning to address data integrity issues, but improving the accuracy and trustworthiness of national broadband maps will take time. Until then, official counts of underserved Americans should be treated with caution.

(Photo by Isaac Smith on Unsplash)