‘AI hype’ risks clouding common sense
Written by Peter Walker
The current hype around artificial intelligence (AI) is getting to proportions where it risks overtaking common sense in terms of business decision-making, according to several experts.
Speaking at the IP Expo 2018 conference yesterday afternoon, Sophos senior security advisor Paul Ducklin told the audience that senior executives should avoid “letting AI hype take over your common sense”.
He continued: “AI can be the hammer that board members use to run around searching for nails in their business… don’t invest just because you think everyone’s going to look down at you if you don’t.”
However, Greg Benson, chief scientist at SnapLogic, told delegates during the same panel session that AI can be a competitive advantage. “It needs to be on any company’s mind from the beginning, because if you’re not thinking about potential applications, then your competitors will be.”
The conversation moved on to the idea of ‘AI washing’, or companies claiming to be driven by AI, when in reality they are not at all, or not as much as they are letting on.
Daniel Yin, head of innovation for EMEA at RingCentral, explained that it’s certainly a good marketing tool to be able to play up your AI innovation to create interest. He noted a recent announcement from Marks & Spencer about automating parts of its call centre operations, which while technically was AI-based, was in fact “a very simple use case”.
Benson said that aspects of this AI washing could be seen as positive though, as “AI is such an important software tool that we need a call to arms, a whole new movement to create change”.
The panel was in agreement though that at the moment, AI is still fairly limited in its useful application. Yin suggested that the technology is best applied to things like optimising information from large data sets, giving the example of speech analysis in contact centres to work out caller sentiment.
“You really need to understand what AI is good and bad for before investing,” he commented.
Ducklin added that within cyber security, some startups were guilty of overhyping their AI-driven products, which in reality only gives a solution in one area. “Just because you ticked that one box, using machine learning to discriminate between good and bad, doesn’t mean you’ve solved the whole problem.”