Enterprise IT shops are finding that artificial intelligence isn't just for robotics anymore.
Science fiction and blockbuster movies may portray artificial intelligence, or A.I., technologies as the brains running robots trying to take over the human race, but the technology today is being used for far more benign purposes.
While A.I. is being used in smartphones and self-driving cars, it's also working its way into the enterprise to filter spam out of email, handle complicated scheduling or detecting fraud in big data deployments.
"I think IT probably needs artificial intelligence," said Stephen Smith, a professor who specializes in robotics and A.I. at Carnegie Mellon University. "There are increasing cyberattacks we're dealing with. We have bigger, more complex problems with all the issues arising out of the explosion of the Web and all of our big data. We're already using A.I. It's already there. I think A.I. in the enterprise is going to start to cascade."
At last month's AAAI-15 conference in Austin, Texas, participants from Xerox Corp., Ford Motor Co. and NASA's Ames Research Center demonstrated the artificial intelligence applications that their organizations are using.
Artificial intelligence is about making intelligent computer systems, such as robots or software that handles financials data, that can learn as they go and handle tasks that traditionally required people to get the work done.
Despite fears about the development of A.I. and its sci-fi reputation, the technology isn't new to the enterprise. It's simply been used under different names, such as email filtering or speech recognition, instead of the umbrella term of A.I. The technology is also appears to be enjoying a growth spurt in the enterprise as executives and IT managers find artificial intelligence well suited to taking on increasingly complex business problems.
Google uses A.I. in search. The iPhone's digital assistant Siri is based on artificial intelligence, as is Internet of Things technology that is starting to be used in homes.
A.I. researchers and industry analysts say these examples are just early steps in using A.I. in the enterprise.
For instance, artificial intelligence is being used to prioritize email, for planning and scheduling, culling through big data, speech recognition and security.
"Without a lot of headlines, it's everywhere from customer service and better routing to finance and fraud detection," said Oren Etzioni, CEO of the Allen Institute for Artificial Intelligence. "It's actually inevitable that we will have more A.I. in the enterprise because of several trends coming together, including more compute power, desire for productivity growth and the big data we're dealing with."
Enterprises are increasingly turning to artificial intelligence because of the explosion of data they have to manage.
"The notion that a human analyst can look at all of this data unaided becomes more and more implausible," Etzioni said. "You can't have a person sitting there watching Twitter to protect your brand. There are just too many tweets. You need help protecting your brand. You need tools. You need A.I. tools."
Xerox PARC and A.I.
Xerox's well-known PARC lab uses artificial intelligence to detect fraud and abuse in health care data.
Eric Bier, a principal scientist at PARC (Palo Alto Research Center), which is known for major roles in the development of the Ethernet and laser printing, talked to an audience at the AAAI-15 conference about its Xerox Program Integrity Validator or the XPIV system.
Bier, and the XPIV team, are working with the federal government to find fraud in areas such as Medicare, the government's health insurance program for those 65 and older.
"There's a tremendous amount of data to go through," Bier said. "The old phrase of 'looking for a needle in a haystack' comes to mind, but then at least you know what a needle looks like. People are clever. You need algorithms that learn from the data and find behaviors that are odd that will call someone's attention to something funny going on."
The PARC A.I. program culls through enormous data sets looking for patterns that signal what could be suspicious behaviors. For instance, the program looks for repeated instances of patients passing by several pharmacies to go to one on the other side of town, patients who buy nothing but narcotics or patients who continually use different pharmacies.
"We're using a visual analysis of the data," Bier said. "The technology learns from the data. We try to focus on making a system that's fast. A lot of the analysts we work with have had bad experiences with systems that might have needed hours or days to get a report. We like to focus on getting answers in seconds."
Ford uses A.I. for scheduling
At Ford Motor Co., managers are using an A.I. system to take on what was becoming a massive, and time-consuming, job of scheduling new hires through a three-year training program.
Instead of a group of people spending hours trying to organize a growing number of program participants, job requests and assignments, the A.I. system – built, coincidentally, by one of the new hires – handles it for them.
"It was [a problem] that was taking away time from people who didn't have time," said Leonard Kinnaird-Heether, an A.I. researcher at Ford who built the program. "A.I. was a good idea for this because this problem represents a core function that A.I. can take care of ... We developed a tool to automate it so we can give that time back."
Google search and A.I.
While some companies use artificial intelligence for individual projects, Google is focused on building A.I. into its core search technology.
"How important is A.I. to Google? Very," said Geoffrey Hinton, a distinguished researcher at Google and a professor at the University of Toronto. "A lot of the problems Google needs to solve -- to bring you the stuff you want to know about -- are A.I. problems… In search, understanding what a document is saying is going to give you much better retrieval. Until a few years ago, image retrieval was done using text that described the image. To better retrieve images for you, the system needs to understand what the images are about -- what's in the images."
For that, Google relies on artificial intelligence.
According to Hinton, that intense focus on A.I. research at companies like Google, and institutions like Carnegie Mellon, is resulting in siginificant progress in the field.
"Things are looking very positive," he told Computerworld. "We're beginning to solve problems a few years ago we couldn't' solve, like recognizing images. We're suddenly getting better at machine translation. That's had a big impact on speech recognition and object recognition. That is closer to the core of Google because it involves understanding sentences, and if you can understand what a document is saying, you can do a much better search."
Lynne Parker, a professor at the University of Tennessee and a division director in Information and Intelligent Systems with the National Science Foundation, said many people still think artificial intelligence is a future technology, but they're missing out on how integrated A.I. has become in the enterprise.
"For most companies, if there's a technology that will help them do their jobs better and more efficiently, they just want it to be reliable, dependable and understandable. My guess is A.I. is not a scary term any more."
Carnegie Mellon's Smith said the more enterprises delve into A.I. technology, the more A.I. will lose its mystique.
"That tech transition has always been a challenge. You kind of have to move in steps," Smith said. "As users get more comfortable with these systems and see the decisions they're making, they'll see they're not doing crazy things, and you'll have the opportunity for better plans and better scheduling and you'll be more productive."