TAC Working Groups Eyeing Recommendations on AI, Spectrum Sharing
The FCC Technology Advisory Council’s working groups will likely propose that the council issue recommendations about AI, spectrum sharing and propagation modeling, according to presentations at Thursday’s TAC meeting. During the meeting, TAC Chairman Dean Brenner said he will follow FCC Chairwoman Jessica Rosenworcel’s lead and leave his post Jan. 20. FCC Commissioner Brendan Carr, the incoming chair, should “pick the person that he wants to lead the TAC,” said Aira Technologies' Brenner, who has chaired TAC for three years. The TAC's charter expires in September.
Sign up for a free preview to unlock the rest of this article
Communications Daily is required reading for senior executives at top telecom corporations, law firms, lobbying organizations, associations and government agencies (including the FCC). Join them today!
Most of the TAC’s working groups are in the preliminary stages of preparing “charter items” for the full TAC to vote on as possible recommendations for the FCC, but many of the groups previewed preliminary versions at Thursday’s meeting. On frameworks for advanced spectrum sharing, the agency should create a multistakeholder group of federal and commercial representatives to reach consensus on technical requirements, said Nokia Bell Labs Principal Spectrum Standardization Leader Amit Mukhopadhyay, who led the effort on that item. The FCC should also continue keeping spectrum rulemakings technology-agnostic but also ensure users continue to improve their technology, Mukhopadhyay said.
The TAC’s subgroup on spectrum repurposing wants to counsel the FCC to consider opportunities to free the upper C-band to at least 4.1 GHz, said CTIA CTO Tom Sawanobori. The agency should evaluate the impact of further repacking of satellite services and learn lessons from the previous repurposing of the C-band, Sawanobori said. Testing interference with aviation equipment should have happened earlier in that process, he said. The FCC should also monitor evaluations of hybrid sharing between Wi-Fi and unlicensed spectrum in other countries and consider the implications for similar sharing in the U.S., Sawanobori said.
AI applications make heavy use of connections and networks, and that means the FCC will have a big role in the technology's advancement, said Lisa Guess, co-chair of the TAC’s AI and Machine Learning Working Group. The AI group unveiled a host of preliminary recommendations for AI policy, including that the FCC develop a telecommunications-specific large language model (LLM) and serve as a neutral “convener,” gathering entities to guide interoperability and the development of network architectures. The FCC should also prepare to develop testing frameworks and best practices for AI and machine learning components of networks, and it should hold multistakeholder meetings on adjusting rules for wireless systems to account for AI, the working group said.
“The wide use of AI/ML and various forms of software” could “change the way the FCC eventually poses its rules and regulations for wireless systems, replacing hard formulas and radiated power limits with statements of behaviors and intent,” said one of the AI working groups preliminary recommendations.
The TAC subgroup studying spectrum for low-power and indoor uses wants the FCC to adopt a “containment-centric” approach to such spectrum, said Kyndryl Vice President Jason Jackson. The agency should define contained environments, set measurable standards for them and leverage tech such as RF shielding and dynamic spectrum access. The subgroup looking at ways to improve propagation modeling wants the agency to continue refining the effectiveness of such modeling used in spectrum sharing by incorporating more realistic techniques, said Motorola’s David Gurney. There is “always a delicate balance between making more spectrum available and protecting incumbent services,” he said. Gurney also said the group wants to recommend that the agency continue incorporating clutter modeling techniques to increase the accuracy of propagation modeling.