Mobile Experts has released research on how artificial intelligence and machine learning (AI and ML) will impact the 5G network.

The group looked at the issue through three lenses: The impact of AI assistant applications, the use of AI and ML to optimize 5G networks, and modeled 5G network demand compared with availability in dense urban areas.

The impact of AI assistant applications initially will be minimal because use cases such as ChatGPT rely on simple text or images. However, the future use of AI and augmented reality is expected to drive “significantly” more traffic, particularly through the use of video, some of which will feature high frame rates and resolutions. This could lead to capacity issues on urban 5G networks if high frame rates and resolutions are used.

The potential use of AI and ML to optimize 5G network performance could moderate some of the challenges. Mobile Experts separated the impact of real time and non-real time AI models and analyzed details such as channel state estimation, beam coordination, traffic steering, and other algorithms to quantify the amount of capacity that would be gained and potential revenue in software, training, computer hardware, and IP licenses.

Finally, the researchers “put both sides together” and separately modeled uplink and downlink demand compared with a 5G model. The Mobile Experts’ finding was that AI assistants and augmented reality will strain the 5G network, but AI will create more capacity than consumers will use through 2030 and perhaps beyond. The variable is video analysis in automotive, drone, and industrial applications.

“New applications are coming to your smartphone, taking advantage of on-device AI processing capability,” Mobile Experts’ Principal Analyst Joe Madden said in a press release about the 5G network research. “We have been able to identify which apps will run on your phone, and which ones will require network connectivity. The frame rates, resolution, and persistence of video inputs will have a profound impact on the 5G network.”