“On any shortlist I’m considered to be one of the world’s leading experts on applied artificial intelligence,” said JT Kostman at a NetApp breakfast in Sydney yesterday.
The American’s claim has substance. Kostman is a former Time Inc chief data officer, was once Samsung Electronics’ chief data scientist, and says he has advised the US Department of Homeland Security, the FBI and the Department of Defense on AI and analytics.
But even he isn’t quite sure exactly what AI is.
“I don’t think anybody really does know what artificial intelligence is. I’m not sure that I’m really sure, what it really means. My definition of AI, truly, I define artificial intelligence as getting computers to do stuff that they do on TV and in the movies. That’s really a great definition of AI. That’s what AI really is,” he says.
“It’s about getting computers to do what computers do on Star Trek. Hopefully not what they do on The Matrix or Terminator movies. It’s getting them to understand us, to do what we want, to do the kinds of things we used to think only we could do,” Kostman explained.
The general lack of understanding and expertise in AI was putting the power of the technology in the hands of too few companies, Kostman said.
He pointed to a recent study by Montreal start-up Element AI which estimated there were fewer than 10,000 people in the world with the expertise needed to create machine learning systems (a similar study by Tencent put the number at 300,000).
They were all being employed on huge salaries by a handful of companies – the likes of Google, Facebook, Alibaba and Tencent – Kostman said.
“There aren’t many people who know what they’re doing who are able to stand up and transform organisations using these practices. And these companies are buying them all up. Of those 10,000 people they’re trying to acquire all that talent and they’re leaving you in the dark,” he said.
Confusion around AI was also giving rise to a “cottage industry” of “pretenders and snakeoil salesmen”.
His comments are supported by Gartner research that found vendors are over-egging the AI capabilities of their products to cash in to increased industry interest.
“They don’t know what they’re talking about they don’t know what they’re doing. We have people who are MBAs from, frankly, third tier colleges who have read a few copies of Wired and HBR, they’ve learned the lingo and the language and they can do the buzzword bingo and they come in and say I’m an AI guru and they’re full of crap and people are suffering as a consequences. Organisations are folding as a consequence,” Kostman said.
To avoid being fooled, executives should arm themselves with at least a basic understanding of both coding and machine learning, Kostman said. Once they had, there were things businesses could do by themselves in the space, even if their companies lacked the might of say JP Morgan.
“It bothers the hell out of me that a handful of companies will rule the world and you will be on the sidelines and screwed. It doesn’t seem right to me,” he told the audience.
However, executives at companies above a certain size should defer AI initiatives to those that know what they are doing, Kostman added. Easy tools like those offered by Amazon Web Services, IBM and Microsoft didn’t always cut it.
“If you fall on the way home you can probably put on a band aid – I don’t recommend doing surgery. And that’s the distinction,” he said.
“What you really need ultimately is an AI Sherpa, someone who can help you see the way."