The technology driving so many software applications today is built on artificial intelligence, and Seattle-area companies are right in the middle of those developments.

Share story

Chances are the entity managing your favorite smartphone app or Internet service isn’t a person.

Algorithms are setting the price of your airline ticket and hailing your Uber driver. They’re placing the vast majority of stock-market trades.

And we’re only at the beginning of a transition that is going to make the algorithms behind the software people interact with better able to understand and react to humans, technologists at a gathering of Seattle’s burgeoning artificial-intelligence industry said Wednesday.

“Every application that is going to get built, starting today and into the future, is going to be an intelligent app,” said S. “Soma” Somasegar, a venture partner with Madrona Venture Group and a former Microsoft executive.

The event, hosted by Somasegar’s Seattle-based venture-capital firm, was held to highlight the cluster of companies in the region working on the cutting edge of intelligent software, including in the discipline dubbed machine learning.

The concentration of work on artificial intelligence here has its roots in the area’s software-development talent, as well as research at the University of Washington. Among the speakers at Wednesday’s event was Oren Etzioni, artificial-intelligence researcher at UW since 1991 and the chief executive of Paul Allen’s Institute for Artificial Intelligence.

Among the area’s tech giants, Microsoft and have both built voice-activated personal assistants, as well as intelligent tools for developers who plug in to their cloud-computing platforms. Other companies are plugging away at everything from software that better understands human speech to tools that make sense of newly digitized corporate data.

Alexa, Amazon’s voice-powered digital concierge, represents for many consumers the real-life implementation of artificial intelligence. The intelligence technology, hosted in Amazon’s cloud, learns as it gathers data, said Rohit Prasad, the Amazon executive in charge of Alexa. “It gets better every day,” he said.

Bringing it to life, however, posed many technical challenges, such as developing robust signal-processing software and a way for Alexa to understand natural language. Work is still ongoing to make it respond more naturally to users’ requests. “This is a pretty big task in general,” Prasad said.

At Redfin, algorithms help real-estate agents more efficiently fill their schedule with home tours, said Bridget Frey, who leads software engineering at the Seattle-based online real-estate brokerage. Redfin’s data-crunching and pattern-recognition software also churn out which homes a house-seeker might be interested in.

“We’re actually better at finding homes (customers) might be interested in than they are,” she said.

Increasingly, it’s the role of technologists to build and manage such intelligent algorithms, said Joseph Sirosh, a corporate vice president with Microsoft who oversees the company’s data group. Software tools already automate the selection and display of advertisements that accompany search-engine results, for example.

“We live in a time of great change in computing,” Sirosh said. “If you don’t climb on this bandwagon … you won’t be very effective.”

Executives didn’t tiptoe around the ethical concerns coming as researchers march from clever pattern-recognition tools and toward true artificial intelligence.

Philip Cohen, a vice president with Bellevue voice-recognition software maker VoiceBox, raised a scenario that could accompany intelligent agents like the software companion character in the film “Her.”

Say that future version of Apple’s Siri or Microsoft’s Cortana concluded through emotion recognition and search history that a member of a household was contemplating suicide: Should that software be programmed to alert others? What ethical guidelines should it operate under?

“It’s not a trivial question,” Cohen said. “It’s coming, and it’s coming in the next five years.”

Executives also addressed the likelihood that artificial intelligence and further workplace automation will eliminate whole categories of careers.

Advances in automated cars put trucking in the crosshairs, said Kristina Bergman, chief executive of Integris, a Seattle maker of data-privacy software. “The economy is completely changing to a knowledge-worker-based economy,” she said.

VoiceBox’s Cohen said governments may have to take steps to protect jobs should technology displace workers faster than society can create new types of work. As an example, he mentioned Oregon’s law requiring gas-station attendants, and not drivers, to fill fuel tanks.

“That’s not a very efficient way to allocate resources,” Kenny Daniel, founder of algorithm marketplace Algorithmia, interjected.

“But you have people who exist, whether it’s efficient or not,” Cohen said.

Etzioni, the UW and Paul Allen researcher, dismissed fears, peddled in science fiction, of an AI-engineered apocalypse, saying they were ridiculous, particularly given the nascent state of artificial intelligence.

He said the current applications of intelligent algorithms — limited to replicating tasks of a complexity a 5-year-old could master — were relatively primitive compared with full, human-style artificial intelligence.

Pedro Domingos, another UW professor and AI luminary, compared the field’s youth to the state of physics four centuries ago.

“We’re in the Galileo stage of machine learning,” he said. “We’re still waiting for Newton.”

Still, Domingos added, “We’re out of the dark ages.”