Some Have Low Expectations for FCC AI Forum

Industry observers expect the FCC's Thursday forum on AI to be mostly an information gathering session as the agency focuses on one of the hot topics of the year but said questions remain about how large a role, or even what role, the FCC could play in rapidly emerging field. The AI forum is the agency's first since 2018 (see 1811300051).

Comments were due Friday, meanwhile, on a request for information on national priorities for AI, at the Office of Science and Technology Policy (see 2307100060). Several telecom companies and associations said Monday they didn’t file in that proceeding.

The FCC forum will take on two topics, “AI’s Dramatic Impact on Communications Networks and Technologies” and “AI as a Tool and Challenge for Consumer Empowerment,” based on the agenda posted by the FCC. Opening speakers are FCC Chairwoman Jessica Rosenworcel, Commissioner Nathan Simington and Sethuraman Panchanathan, National Science Foundation director. It has a keynote by Margaret Martonosi, assistant director-NSF Directorate for Computer and Information Science. The forum starts at 9 a.m. at FCC headquarters.

There have been a few hints on where the agency could be headed. “We’re doing some work behind the scenes” on AI, Rosenworcel said after the May FCC meeting (see 2305180052). The FCC is looking at AI for “everything from robocalls to network reliability and self-healing” networks, she said then. The agency’s Technological Advisory Council also focused on AI and 6G (see 2307030020).

“It looks like it is going to be a discussion to gather both information and perspective from stakeholders,” said Monica Paolini, consultant at Senza Fili. “I would imagine, and hope, that they are also working on the topic internally, and in fact that they are working on the topic in a wider context. The topics for the meeting are quite interesting, but there is much more to talk about.” Paolini said the experts speaking are the ones to be expected at that type of FCC forum.

"Many agencies are looking into what their piece of the AI regulatory mosaic will be,” said Cooley’s Robert McDowell. The FCC could “try to make the companies it regulates adhere to AI guidelines if they want to avoid enforcement proceedings and have their licenses renewed,” he said: “But let's see how it handles this workshop first."

Jonathan Cannon, R Street fellow-technology and innovation and a former acting legal adviser to Simington, has low expectations. “AI is the hot topic and everyone wants a piece of the pie,” he said: “Agencies are champing at the bit to have an opportunity to engage on AI regulations. I don’t see any connection between AI and FCC.”

There's a lot of opportunity for algorithmic tools “to help with network reliability or self-healing/repair of networks, [but] this doesn't necessitate FCC involvement for carriers to benefit from these technologies,” Cannon said.

The FCC’s look at AI isn’t a surprise, emailed consultant John Strand of Strand. “Notably it seems like a bipartisan undertaking, and indeed AI’s role in comms networks is noted as a theme as well as consumer insights,” he said: “The first thing that comes to mind is privacy and how the FCC would interpret that under U.S. telecom privacy rules.”

There have been demonstrations of how AI “can facilitate much more intensive use of spectrum, and uses against robocalls seem promising,” said Joe Kane, Information Technology and Innovation Foundation director-broadband and spectrum policy. The FCC should focus on ways “to enable the best uses of AI to improve areas under its direct jurisdiction rather than indulging in mission creep to claim regulatory oversight that will hamstring the net benefits of AI throughout the economy,” he said.

"AI has become a hot topic in the news and press due to generative models like Chat GPT, but many use cases exist beyond what we hear in the press,” emailed Jeffrey Westling, American Action Forum director-technology and innovation policy. Modern networks “can use AI and machine learning to improve network optimization and direct traffic more efficiently, lowering latency and service disruptions,” he said: “To the extent that firms use AI and machine learning in areas already regulated by the FCC, the agency will undoubtedly have a role to play.”

If the government decides regulation is necessary for AI, it should adopt a “risk-based approach,” USTelecom said in a filing at OSTP. The comments don’t mention a possible role for the FCC. “The telecom industry understands the importance of trustworthy AI that is safe, effective, ethical, and legal,” USTelecom said: “In pursuing trustworthy AI, policymakers should avoid a prescriptive approach to regulation that could stifle innovation while failing to keep pace with the dynamic nature of a technology that continues to rapidly evolve.”

Public Knowledge urged an expert AI regulator. “While existing sector specific regulators are well-suited to taking on existing and emerging harms in their own domains, AI technology is too important, wide-ranging, and technically complex to rely on disjointed regulatory efforts,” PK said: “Congress must either dramatically expand the jurisdiction of an existing agency, or create a new agency specifically charged to regulate AI.”