What can be done for the 23 million households that relied on the Affordable Connectivity Program? The Federal Communications Commission could start a proceeding to increase the amount of money in the USF so that it can fund an ACP-like subsidy. It has the power under the Communications Act, it has a consistent funding source, and it has a reason with the ACP’s expiration—an event former FCC Chief of Staff Blair Levin describes as the “biggest step any country has ever taken to widen, rather than close, its digital divide.” Given all this, why doesn’t the FCC act? Two words: contribution reform. You see, the USF has one fundamental flaw: it’s funded solely through that fee on phone service. This is a problem because people are shifting from phone-based communications to broadband-based communications, and as they do, phone-based revenue for USF is drying up. Thus, the USF is barely able to support its existing programs, much less take on an ambitious project like funding a $30-a-month broadband subsidy. The FCC could fix this problem by applying the USF fee to broadband services. Doing this would increase the USF’s revenue base and allow it to address the looming ACP emergency.  But it is also controversial—particularly to cable companies who pay very little into the fund. Unfortunately, the FCC largely shut down this option a few weeks ago when it chose to forbear from applying USF fees to broadband providers.

[Gigi Sohn is Executive Director of the American Association for Public Broadband and Benton Senior Fellow and Public Advocate]

Vermont Community Broadband Board has awarded a $20.2 million construction grant to Communications Union District (CUD) Northwest Fiberworx. The grant was awarded at VCBB’s May meeting and comes from the Broadband Construction Program, a State of Vermont grant funded with American Rescue Plan Act dollars. The grant will allow construction to start on the fiber internet network that will serve 22 member towns in northwest Vermont. This $20.2 million grant will allow Northwest Fiberworx to begin Phase 1 of their construction model and serve the region’s least connected areas, reaching more than 3,800 unserved and underserved addresses. VCBB has awarded more than $206 million in grants for the broadband buildout in Vermont. Once Northwest Fiberworx starts construction, nine of the state’s ten CUDs will be in construction and/or actively connecting customers.

This episode of The Divide features Brandy Reitter, executive director for Colorado's broadband office. The state of Colorado recently received NTIA approval on volume two of its initial proposal for the Broadband Equity Access and Deployment program, or BEAD, getting the state one step closer to distributing $826 million for broadband deployment. (Check out both volumes of Colorado's approved BEAD proposal here.) We discuss what that NTIA approval means and next steps for the state's BEAD program (according to Reitter, Colorado expects to launch its sub-grantee selection process in August or September). We also discuss ongoing challenges to access and adoption, including the significant impact of losing the Affordable Connectivity Program (ACP). 

Investments in high-speed broadband have received considerable attention from policymakers and researchers. Governments are committing to increasing available internet connection speeds through massive public investments. These policies are motivated by arguments according to which improvements in broadband connections play an important role in fostering productivity, economic growth, innovation, and knowledge. Despite the ubiquitous usage of the internet among children and youth, very little is known about its effects on their academic performance. How advanced broadband internet affects student performance remains an empirical question, and the existing literature on the topic does not provide a conclusive answer. The goal of this paper is to shed some light on this issue using a municipality-level data set on the diffusion of ultra-broadband connections in Italy, matched with the national standardized test scores in Italian language and Mathematics in 2nd, 5th, and 8th grades. Results show that more broadband availability is detrimental to school performance for students in 8th grade, whereas we find no effect on students in the 2nd and 5th grades, most likely because they are too young to engage in distracting online activities.

New York State legislators recently put a stop to personalized social media feeds for the under 18 set, citing mental health harms. The law they passed takes an unusual approach to the challenging task of regulating social media by focusing on algorithms, rather than platforms or specific content. Lawmakers say the algorithms are addictive. Research agrees—mostly. But the way the research explains addictive algorithms and their harms is more of a constellation than a straight line. Social media companies aren’t transparent about how their algorithms work, so researchers have had to study them indirectly. Testimony from former Meta employees tells us that content algorithms are designed to keep kids glued to their phones. What makes feeds tailored to an individual’s preferences so toxic has less to do with personalization and much more to do with how companies hack engagement. Not only are social media algorithms keeping people on for longer, social media platforms may also be exposing teens to certain types of negative content that keeps their attention.

Connecting the dots of recent research suggests a new future for traditional websites:

• Artificial Intelligence (AI)-powered search can provide a full answer to a user’s query 75% of the time without the need for the user to go to a website, according to research by The Atlantic.
• A worldwide survey from the University of Toronto revealed that 22% of ChatGPT users “use it as an alternative to Google.”
• Research firm Gartner forecasts that traffic to the web from search engines will fall 25% by 2026.
• Pew Research found that a quarter of all web pages developed between 2013 and 2023 no longer exist.

The large language models (LLMs) of generative AI that scraped their training data from websites are now using that data to eliminate the need to go to many of those same websites.