CableLabs CEO Phil McKinney dismissed much of the hype surrounding artificial intelligence (AI) as overblown, warning onlookers should take AI announcements with a pinch of salt.

Speaking with Mobile World Live ahead of CES 2018 next week in Las Vegas, McKinney – boss of the US-based industry R&D lab for cable and wireless technologies – argued companies have gotten sloppy with their use of terms in marketing. While scores of projects will be touted as AI-enabled, McKinney said many conflate big data and machine learning capabilities with what would be viewed as true artificial intelligence.

According to McKinney, real AI will be applied first to areas that require “deep vertical expertise”, essentially capturing and aggregating the specialised knowledge of top-level engineers and making it widely available. Examples of this can be seen in the medical field, where AI is used to read X-rays, and in corrective network maintenance, where the knowledge of the best technicians is applied to do predictive diagnostics, he said.

AI in the network
McKinney said this type of AI will become increasingly important as wireless networks evolve to more and more complex architectures. While complexity can be a drag on companies’ ability to innovate, McKinney noted AI presents the opportunity to “capture the knowledge of top network operators and apply it to addressing, improving and operating those networks”.

With the advent of software-defined networking (SDN) and network function virtualisation (NFV), McKinney reported work is already being done to apply AI to resource management. This, he said, can not only help reduce costs and increase network reliability, but also free up resources for companies to dedicate to innovation.

McKinney explained: “The whole purpose of SDN/NFV is to be very dynamic. The challenge is dynamic is great in theory but executing it requires a lot of analysis to make it work. I think that’s where AI comes in and plays a huge role.”