Artificial intelligence isn't an industry so much as a technology poised to transform business across a wide variety of sectors--and probably more than you think.
During a panel discussion Tuesday at the Las Vegas tech trade show CES, a group of A.I. experts talked about which industries are the most ripe for adoption and application of A.I., and why entrepreneurs and consumers alike stand to benefit significantly from the technology.
The number one industry set be transformed by A.I. appears to be healthcare, with $400 million invested by health care companies in the technology as of last year, a figure that's projected to grow to $3 billion or more by 2020, according to data from the Beacon Center for the Study of Evolution in Action. The retail industry is close behind, with $100 million invested as of 2015, expected to reach $1.9 billion by 2020. Panelists pointed to manufacturing, financial services and government as the three followers to healthcare and retail.
The majority of the activity around harnessing the technology is focused on bringing A.I. to businesses, not consumers, according to panelist Gayle Sheppard, general manager for Saffron Technology within Intel Corporation's new devices group, but consumer-oriented A.I. is expected to grow in the future.
"Whether it's expressed through wearable devices that we have on us or in another broad base of applications, I think it's got a very bright future," Sheppard said, adding that the rise of Internet of Things devices will also drive consumer-oriented A.I. "It's a brand new industry that's just emerging."
So how exactly will A.I. have a transformative impact on business?
Debi Mishra, partner director of engineering and machine learning at Microsoft Corporation, pointed to the example of GE's aircraft engine business, which shifted its business model from selling machinery to selling engines as a service.
"The primitive maintenance prediction of these aircraft engines and the data on these engines is an incredible intellectual property," Mishra said. "You take the data and analytics it generates, you take care of the maintenance, and all of that is something you can package, and then you are selling that as a service." As more IoT devices begin popping up in places like homes and cars, he said, other industries will undergo similar changes to their business models.
For Babak Hodjat, co-founder and chief scientist at Sentient Technologies, automated design is an area within A.I. that could fundamentally change e-commerce. Apparel companies, for example, could one day create new products based on customers' purchasing history and design preferences.
"Imagine a world in which as you interact with what's available in the inventory on an e-commerce site, you're telling the commerce site what you're interested in," Hodjat said. "By virtue of that interaction, you're actually enabling the system to come up with a design that can be 3-D printed and be completely custom designed for you."
Hodjat also cited energy management as one of the industries for which A.I. will make significant advances in the near future. Programming when electric car chargers start charging and how home devices make use of energy from solar panels versus the grid are two examples he mentioned.
"There are fascinating application areas there, not just in discovering patterns but actually in decision making in the homes, like when to turn what on and how to make use of this energy so that it's optimized--not just for the consumer, but for the power plant and energy companies as well."
One area in which A.I. experts have yet to make significant progress is in bringing automation to companies that perform audits of other businesses.
"It's not that you audit a bunch of things and find flaws--you've got to be able to explain exactly why," said Microsoft's Mishra. "It's pretty reasonable to expect forecasting in inventory or forecasting a traffic pattern, but once it gets into much more intellectual subjects of human learning, [it becomes more difficult]."
Still, Sentient's Hodjat warned against assigning limitations on what kinds of problems A.I. has the potential to fix.
"You'd be surprised at what A.I. can do," he said. "A lot of problems that we think are impossible to solve--and are within the exclusive domain of human intelligence--can be solved."