How 5G and artificial intelligence may influence each other

Moving past 4G

The ultimate goal of 5G is to allow for more devices on the network, decrease latency and allow for further innovation through faster network data processing speeds.

The current generation of wireless connectivity, 4G, has reached its capacity and organizations looking to handle the continuously rising level of data must turn to the next chapter. As Dan Bieler, principal analyst at Forrester, sees it, current wireless networks are seeking to expand beyond the restraints of 4G capabilities.  

"There is only a certain amount of data, [and] of devices that can be tracked in a given geographic location by existing technologies," Bieler said. "A few thousand devices per square kilometer, for instance."

This becomes more of an issue as enterprises look to increase the amount of data and number of sensors supported by the wireless networks they use, especially in the industrial sector that requires millions of data points and individual devices. Bieler said any organization that is seeking to significantly increase the number of devices they operate will soon hit a wall when using 4G -- one reason why 5G has been created.  

"5G has a very high capacity of sending data, just the sheer volume and bandwidth that is available," Bieler said. "5G is defined in a way that it supports up to a million devices in one square kilometer."

Combining 5G and AI

Though 5G is incredibly new and still in pre-deployment in many places, and AI has reached widespread adoption, analysts predict the two will flourish together. Artificial intelligence demands high data volumes to train models, and 5G may help deliver these large training data sets more quickly than current methods. With the network and data advantages that comes with 5G, the cemented use cases of AI -- machine learning and in-depth data analysis -- could be improved and expanded.

"5G makes AI better by having the low-latency, high data speeds and also high data device density so you can have all sorts of connected devices," Jeff Loucks, executive director at Deloitte's Center for Technology, Media & Telecommunications, said.

Experts believe that AI models will improve with the speed, decreased latency and increase in capacity that 5G can provide.

"You're able to get into multi-gigabit speeds, you're able to get into latency that's operating at sub-10 milliseconds," Jack Fritz, principal for technology, media and telecommunications at Deloitte, said. "And you're able to really improve reliability."

It isn't just that 5G could bring upgrades to AI, but the opposite is true as well. AI can help route traffic and balance loads in the 5G network.

"AI makes 5G better in [regards to] network performance, simplifying deployments and enhancing service quality," Loucks said.

But 5G remains a young technology that still has yet to be rolled in out in many parts of the country, which means many are still trying to understand its capabilities.

"We are just at an early stage here because 5G has not been fully rolled out to the extent that all the functions and features … are available now," Bieler said.

The intelligent edge

Where these technologies will interact is mainly at the edge of the network, by utilizing a combination of processing power, artificial intelligence and advanced connectivity like 5G, Loucks said.

The intelligent edge is defined by devices distributed throughout a network and close to the users or appliances creating data that have the benefits of AI processing and 5G connectivity, allowing them to function more independently and efficiently.

"AI really helps in delivering the ultimate performance on the devices and certainly plays a significant role in network management as well as network deployment and planning," Fritz said.

By adding AI and 5G to the network, the possibilities for different use cases increase as well as the efficiency of current processes.

"This makes certain usage scenarios possible that traditionally would not have been possible, like a smart port for instance, or like a fully automated manufacturing environment," Bieler said.