AI’s Impact on 5G Network Governance

The growing complexity of telecom networks across the world following the launch of 5G services and the multitude of use cases that it has generated could be a cause for concern. However, experts argue that the advent of artificial intelligence (AI) into the mix could be a stepping stone towards managing these complexities. 

For starters, AI could be handy when it comes to managing network resources better, more specifically those around the electricity that is required to manage a cell site. Of course, one would have to factor in that AI itself could become a power guzzler if mechanisms aren’t pre-installed to contain it. 

Optimizing the power consumption at cellular level

One such use case could be optimizing power by switching on and off some cells on antennas during lean periods such as night time. In fact, AI and machine learning could come in handy to save power even when calls are on as most conversations do have pauses. There can also be time slots in a 24-hour cycle when communication may not exist. 

According to Subhankar Pai of Capgemini Engineering, such use cases may not indicate immediate or significant benefits but considering the volumes of large cell sites on a network, the cumulative impact could be considerable. Large telecom carriers could be spending as much as $1.5 billion a year on energy, he told SdXCentral. 

Pai believes that operators could be running large cell sites with ongoing deployments of smaller ones for more targeted and localized coverage. Keeping all of them going at all times could be a massive energy drain that adds significant costs. By a 10% reduction in energy use, telecom providers will make large savings while also achieving net-zero targets. 

In fact, experts around the world now believe that radios that aren’t being used need to be shut off and if someone starts using it, the system just fires back up, Addressing a 6G Summit at Brooklyn recently, Chris Sambar of AT&T says such a system could potentially save power equivalent to 14,000 homes in a year. 

However, some challenges need to be addressed

However, there could be some data quality challenges when it comes to using AI around 5G networks, which automatically calls for machine learning to generate better AI models. Telecom operators may find it tough to utilize such data that they garner but are loath to share with external sources for reasons of compliance and competition. 

Would operators share user information such as mobile numbers or usage patterns and locations with an external agency? For starters, legal issues may prevent them from doing so and if this cannot happen, the quality of data and predictions thereof would remain dodgy. Of course, one could anonymize data, but that can happen only if telecom operators unify to do it.

Another imponderable at this juncture relates to the location of the services and the storage of data not necessarily being in the same geography. Take the case in India where the government has allowed data to be stored outside the country with the sole exception of locations that are considered “not friendly”. 

In spite of these challenges, the fact remains that mobile telephony in the 5G era and the future would only get more complicated as businesses develop smarter use cases, both for individuals as well as enterprises. And in doing so, they will require high volumes of computing that could only add to the pressures on net-zero targets and push up overall costs.