Click to Skip Ad
Closing in...

Is Google crossing a line with new AI that pretends to be human?

Published May 9th, 2018 11:19AM EDT
Google IO 2018
Image: Dmitry Kuznitsov, Dreamstime

If you buy through a BGR link, we may earn an affiliate commission, helping support our expert product labs.

Most people, even if they don’t know anything about the norms of journalism, are aware that reporters are expected to identify themselves as such to sources. That there’s something … shady, let’s say, about a reporter attempting to get information out of someone by pretending to be an ordinary member of the public.

The dynamic instantly changes, as it should, when someone realizes they’re dealing with a member of the Fourth Estate. Which is one reason why Google’s creepily human-sounding assistant Duplex it unveiled Tuesday at its I/O developer conference — an assistant that’s complete with what Google calls “speech disfluencies” like the hmms and umms of normal human conversation — has already been the focus of so much attention and headlines.

Bottom line: Google has made a leap forward in developing a computing tool that can perform some of the menial tasks of daily life — booking a table at a restaurant, scheduling a haircut — and in the process sound so close to a human doing it that you can’t really tell the difference.

“Give me one second,” a real-world hair salon receptionist said in the demo, after realizing Google’s Duplex — which she apparently thinks is a human on the other end — needs to make an appointment.

“Mmm-hmm,” Duplex responds, generating a round of laughter from the audience.

But stop and consider that. Google has added all of the little oddities of human speech precisely to make you feel like it IS a human on the other end. It wants you to think that, without disabusing you of the notion. But does that cross some sort of line? A line demarcating the bounds of normal interaction and the expectations between parties to that interaction?

Watching the real-world calls with Duplex unfold at I/O inspired tech journalist Steven Levy to sum up the problem that needs to be addressed this way: “Is it ethical to have a human-sounding robot interact with someone without informing the other party that he or she is in conversation with an it?” he tweeted to his followers. “Real question.”

To be sure, in this case where mundane tasks are being performed like making a hair appointment, there may not be much that actually changes if the human party on the other end realizes they’re talking to a computer. The appointment still needs to be made. The human would still need to present some possible times or lay out the parameters of booking and ask for the client’s preferences. Still, the argument goes, people ought to realize what they’re dealing with.

There’s some indication already that Google understands this and that the plan is not to build something where proactively fooling you into thinking it’s human is the goal. Two Google engineers wrote on the company’s AI blog Tuesday that the Duplex technology is “built to sound natural, to make the conversation experience comfortable. It’s important to us that users and businesses have a good experience with this service, and transparency is a key part of that. We want to be clear about the intent of the call so businesses understand the context. We’ll be experimenting with the right approach over the coming months.”

It’s technology that’s simultaneously super cool but potentially — well, terrifying is a word that’s already been used by no less than ex-Googler Chris Messina: “Google Duplex is the most incredible, terrifying thing out of #IO18 so far.”

“Terrifying” also, because, consider: Duplex, according to the Google engineers, is self-aware enough to identify the tasks it won’t be able to complete itself, like scheduling a really complicated appointment. In those cases, it can give what amounts to a digital tap on the shoulder to a human operator to help intervene. (Paging Elon Musk: “Self-aware. I told you so.”)

The demo showed us we’re closer than ever to the world of “Her,” with a Samantha that’s close enough to feeling like a human that we have a ton of new rules about human-computer interactions to consider. Technologists frequently tout idealistic sentiments like a goal being to perfect technology such that it “gets out of the way,” to the point where you can focus on what you need to do without getting frustrated by tricky menus, required gestures and the like. But that doesn’t mean everything needs to go away — like robbing people of an understanding of what’s on the other side of the screen. Or in this case, the phone. Duplex doesn’t identify itself the way a human might — there’s a bot holding for you on line two — but maybe it should.

Andy Meek Trending News Editor

Andy Meek is a reporter based in Memphis who has covered media, entertainment, and culture for over 20 years. His work has appeared in outlets including The Guardian, Forbes, and The Financial Times, and he’s written for BGR since 2015. Andy's coverage includes technology and entertainment, and he has a particular interest in all things streaming.

Over the years, he’s interviewed legendary figures in entertainment and tech that range from Stan Lee to John McAfee, Peter Thiel, and Reed Hastings.