AI Will Be Ubiquitous in 2020 but Overhyped in 2017 – Gartner

Artificial intelligence will be baked into nearly every new software product and service by 2020, but the rush to claim AI integration today is causing confusion, according to a new report from analyst firm Gartner.

While virtually every company across industry verticals is busy today figuring out what AI can do for their business, the hype around it is causing many to jump the gun on building and marketing AI-based products, rather than truly understanding them and their value first, Gartner Inc. says, noting that AI didn't crack its top 100 search terms on gartner.com in January 2016 but was number seven by May 2017. (See Is IBM's Watson Overhyped & Soon to Be Outdone?)

That said, by 2020, the firm expects it to be a top five investment priority for more than 30% of CIOs.

The various forms of AI are attracting equal parts interest -- for their potential to automate processes, simplify complex tasks, improve security and more -- and fear over the potential to displace jobs or even lives. (See Will AI Create More Jobs Than It Destroys?)

For more on artificial intelligence in the world of telecom, visit the dedicated automation content page here on Light Reading.

As it relates to telecom and enterprise, the threats are a little less existential, and AI is seen as augmenting humans' work rather than displacing it. Gartner defines AI as the "systems that change behaviors without being explicitly programmed, based on data collected, usage analysis and other observations." It identifies three things that tech providers need to understand in order to move beyond the hype and make AI work for them:

  • Build a collection of case studies with quantifiable results achieved using AI in order to differentiate from the more than 1,000 vendors that are "AI washing," or applying the AI label to liberally and confusing buyers.

  • Keep it simple and use proven, less complex machine learning capabilities to solve a business challenge rather than using AI just for the sake of it.

  • Offer solutions to actual business problems like process automation rather than focus on cutting-edge technology. Over half of respondents to a Gartner survey also indicated that a lack of staff skills was a top challenge to adopting AI, so Gartner advises that vendors highlight how their solutions address the skills shortage. (See IBM: AI Needs More Than Just Technology.)

    — Sarah Thomas, Circle me on Google+ Follow me on TwitterVisit my LinkedIn profile, Director, Women in Comms

  • James_B_Crawshaw 7/18/2017 | 7:08:32 PM
    AI in telecoms Re AI washing - Andy Tiller wrote in March: "The bar for what counts as Artificial Intelligence is continually rising. Technologies should lose their AI status when they become routine (optical character recognition is no longer recognised as an example of AI for this reason)."

    One of the key applications of AI in the telco industry is in process automation. The current state of play is that operations staff write scripts to automate processes. A number of companies offer Robotic Process Automation solutions which simplify this script-writing. RPA is widely used in financial services while telecom is a relatively late adopter. To take this to the next level we have initiatives like TMF's SID which aim to create device and service models which enable greater programmability of the network. By embedding AI in these orchestration systems and lower level RPA/scripts the hope is that we can increase the level of automation in network operations even further.

    You can read more about this in this article here: Automatic for the Operations People
    Sarah Thomas 7/18/2017 | 5:32:51 PM
    Re: hype cyclin That's fair, but I think that's where the many definitions of AI and machine learning come in. It's not a lie so much as a gray area, or just typical marketing...
    mendyk 7/18/2017 | 5:22:30 PM
    Re: hype cyclin But that's really not hype so much as deception or obfuscation.
    Sarah Thomas 7/18/2017 | 5:18:19 PM
    Re: hype cyclin Similar to how 5G will be a reality in 2020, but there are a ton of pre-5G, 5G-esque, 5G-ready type of deployments happening today. 
    Sarah Thomas 7/18/2017 | 5:18:19 PM
    Re: hype cyclin Similar to how 5G will be a reality in 2020, but there are a ton of pre-5G, 5G-esque, 5G-ready type of deployments happening today. 
    mendyk 7/18/2017 | 4:37:26 PM
    Re: hype cyclin The Gartner analysis as reported here is a bit ... puzzling. How can something be overhyped if the ultimate conclusion is that it will be a major force in three years' time?
    Sarah Thomas 7/18/2017 | 4:02:07 PM
    hype cyclin Gartner's definition of AI is helpful, because there are a lot of definitions floating around out there, which is one reason vendors are busy claiming they have AI solutions when it may not be the true technology. It runs the gamut from basic machine learning to the scary, potentially deadly robots Elon Musk is calling on the government to regulate. 

    The hype, various definitions and bandwagoning seem very familar, because the same thing happened with 5G (and most new technologies)...It'll quiet down, but AI really will be vital in the very near future. 
    Sign In