Artificial intelligence potentially could be used to improve broadband maps, said AI experts at a workshop hosted by the Federal Communications Commission (FCC) and the U.S. National Science Foundation (USNSF) yesterday.
Despite the inevitable challenges that AI brings to the table, it offers huge potential gains for telecommunications companies and their customers, according to a wide array of stakeholders who spoke at the workshop, entitled The Opportunities and Challenges of Artificial Intelligence for Communications Networks and Consumners.
The experts who spoke at the workshop eagerly elaborated about the many opportunities that AI presents for future spectrum management, network resiliency, and data collection/use. They also cited standardization and collaborative efforts that will help to ensure AI’s success for telecom providers and others within the telco ecosystem.
For example, panelists were asked how AI can help the FCC advance or achieve some of its high-level goals such as the pursuit of 100 percent nationwide broadband, network resiliency, the advancement of US global competitiveness and increasing network efficiency.
“My first thought that comes to mind as we are looking at 100 percent broadband coverage is that we need a very accurate national broadband map,” said Lisa Guess, Senior Vice President, Solutions Engineering, Ericsson/Cradlepoint.
“AI can – not by itself, because we have to have the all the necessary data to compile and correlate – but AI can be an important tool to get us to a granular broadband map so we will know when we get to 100 percent.”
Fellow panelist Ness Shroff, Director, NSF AI Institute; Professor of ECE and CSE, The Ohio State University, agreed with Guess.
“Predictive modeling approaches could be used to create a detailed broadband map and AI could be useful for coming up with methods of automating the process of evaluating broadband access,” he added.
An AI-generated and maintained map could be used to depict how broadband access is being used and to discern which programs are bringing the most value to telcos and consumers, he explained.
“AI could be used to automate processes to check on these programs and their success. That could be useful on the back end,” said Shoff.
The FCC has received considerable criticism for the National Broadband Map, even after updating the process for creating the map. Most stakeholders agree that the latest version of the map, which went live several weeks ago, is more accurate than the version that came out six months earlier. But critics say it still needs improvement.
Perhaps the AI recommendations made by the panel could be just what the commission needs to close the remaining gap.
Fake Comments
AI also could be used to streamline the commenting process on rulings proposed by the agency, said Harold Feld, Senior Vice President, Public Knowledge, at yesterday’s workshop.
“Let me add one thing which will seem outside all of this – fake comments,” said Feld.
The FCC is legally bound to accept all public comments on its rulemaking proposals. Considering how many comments the FCC receives and must process individually, fake comments can slow this process considerably. The FCC could use AI to help determine whether the comments it receives are genuine by looking for suspicious patterns, etc., and weeding the fake ones out at the start of the process, he explained.
“This is something the FCC should consider,” he said.
Joan Engebretson contributed to this report
