Artificial intelligence, or AI, seems to be the most buzzed-about topic in technology lately.
It’s helping display the right ads within your social network, recommend the next product to buy from your favorite online retailer, and—soon—direct your autonomous car around town.
In other words, every Fortune 500 executive should be thinking about how best to use AI within his or her own company. Baidu chief scientist Andrew Ng, speaking at a Fortune Brainstorm Tech dinner at The Bellagio in Las Vegas, believes CEOs should go further.
Much further.
“You need a chief AI officer,” Ng told Fortune assistant managing editor Adam Lashinsky. “If you have a lot of data and you want to create value from that data one of the things you might consider is building up an AI team.”
In the old days—say, the age where electricity was the buzzy technology du jour—you’d probably do the same thing. Appointing an executive to deal with what was once a complicated, little-understood technology—a chief electricity officer—would have been a reasonable decision for a CEO to make at the time. Today it’s almost inconceivable.
Artificial intelligence requires a similar approach in the 21st century, Ng said.
“Seek people who can work cross-functionally and have skills to take shiny [new] tech and contextualize it for your business,” Ng urged the executives in the room. “You can’t just download it and bolt it on to an organization.” For this reason, “there’s a talent war for AI.”
He added: “AI is the new electricity”—and lighting was simply its “first killer app” on the way to transforming every industry.
This is why Baidu, the Chinese web services giant and Ng’s employer, is investing so heavily in artificial intelligence technology. AI stands to transform every industry on the planet, Ng said. Better still, there’s a “relatively clear path” for it to do so.
“My friends and I sometimes challenge each other to name an industry that will be difficult for AI to transform,” he said. Maybe cutting a person’s hair—though judging by his own haircut, even that may be possible, Ng joked.
“AI is creating tremendous economic value today,” Ng said, largely due to applications involving the concept of “supervised learning”—the input/output improvements that allow spam filters to become more sophisticated and web services to become more personalized.
The untapped potential of AI is quite clear. Today, Baidu—considered by many to be a leader in AI development—trains its speech system on 50,000 hours of data, or about five years’ worth. “No human child needed five years nonstop of audio with transcription to learn,” Ng said.
But it’s not raw horsepower that will make the most impact. “A lot of the game of AI today is finding the appropriate business context to fit it in,” Ng said. “I love technology. It opens up lots of opportunities. But in the end, technology needs to be contextualized and fit into a business use case.”
And what of concerns that AI development will accelerate to the point where evil machines take over the world? Responding to a mixture of laughter and groans, Ng urged his audience to not get too far ahead of itself.
“Worrying about killer AI robots today is like worrying about overpopulation on Mars,” he said with a grin. “It’s not productive to work on this problem now.”