Entrepreneurs

In offices of the future, everyone will have a souped-up Amazon Echo-type robot assistant

Photo by VCG

Ever have one of those days where you really wish you had a second set of hands? Or maybe an assistant?

In next generation offices, that may just be a reality, because everyone will have a robot helper, says Tolga Kurtoglu, CEO of Silicon Valley research and development company PARC.

"You can think about how we have the Amazon Echoes and Google Homes at home and interact with them. And all of that data and all of that interaction would be delivered to people in a way that those agents understand the workflow, the task, the corporation hierarchy," says Kurtoglu to Recode's Kara Swisher.

Next generation office spaces will "be loaded with machines that can talk to you and that can interact with you," says Kurtoglu, who, prior to joining PARC in 2010, worked as a researcher at NASA.

Critical to this vision is the idea that humans and robots will work together, according to Kurtoglu. And that's the focus of what he does at PARC, a subsidiary of Xerox that does contract innovation work for private companies and the government.

The "really interesting project and technology that we're working on is about how to bring together these AI agents or computation agents and humans in a way that they form sort of collaborative teams to go after tasks," says Kurtoglu.

"What we're talking about here is more of a symbiotic team between an AI agent and a human in a way that they solve the problems together. It's not one of them tells the other one what to do, but they go back and forth. And they can formulate the problem, they can build on each other's ideas and things of that nature."

Kurtoglu's vision is particularly relevant, as Silicon Valley leaders like Elon Musk and Mark Zuckerberg debate the future of AI. Zuckerberg is optimistic that AI will make our lives better. Musk has threatened that AI poses a catastrophic risk to humanity and says Zuckerberg's optimistic perspective is naive.



The way forward: Trust and privacy concerns remain

Kurtoglu says there are barriers to overcome on the path towards widespread partnership between humans and their artificially intelligent robot partners, namely trust and privacy.

For humans to be willing to work easily with robot assistants, the robots need to be able to tell humans how and why they arrived at decisions. Currently, that is not the case.

"We're trying to build trustable AI systems. And one way to get there is for the AI systems to be explainable," says Kurtoglu. "So imagine an AI system that explains itself. So if you're using an AI agent to do medical diagnostics and it comes up with a seemingly unintuitive answer, then the doctor might want to know why, right? Why did you come up with that answer as opposed to something else? Today these systems are pretty much black boxes, so you put the input, it just spits out what the answer is."

Even still, trust will come slowly for humans, says Kurtoglu, because it will require a change in behavior. That does happen, though.

"So it seems a distant past, but I remember, I remember very vividly when I first moved to the U.S. I was trying to do online shopping for travel and I was supposed to get my credit card information. And I was like, 'Should I really trust this?' because it wasn't the common way of booking," says Kurtoglu, who was born and raised in Turkey.

"Then I did it. I think with time, as the adoption curve goes up and whatnot, there's a certain level of trust that builds in, but I think the rate at which these algorithms are advancing now, they're going to be everywhere. I think it's going to be significant."

"We're trying to build trustable AI systems. And one way to get there is for the AI systems to be explainable." -Tolga Kurtoglu, CEO of Silicon Valley innovation shop, PARC

The other issue with working side-by-side with a robot the sees, listens and responds to a human is that the robot will see and listen to all of a human's behavior. That results in a loss of personal privacy and an increase in cybercrime, Kurtoglu says.

"We need to think about the signs in technology that would then make our data, our workflow, more secure and more private as well as the personal data more secure and more private," he says.

As artificial intelligence increasingly becomes a part of the world we live in, Kortoglu says there will always be a place for human intelligence.

"There's still the complexity in the world. There's still those edge cases, if you will, or cases that are very, very complex, where humans are really good intuitively or experience-based to look at the situation at hand and make a judgment, where computers can't."