It's an odd feeling when your own science fiction catches up with you.
A decade away I wrote a short story about a kind of Artificial Intelligence called Verts. The Verts only existed on 2-D screens, but the screens were ubiquitous. They were fake, but CGI had crossed the uncanny valley: They contrived to look more real than their human customers. They were always trying to sell us something — the name derived not from virtual but from advertising — but we didn't mind. They'd conquered the planet by knowing us better than we knew ourselves. It seemed like humanity to create its first self-aware AI in the form of ads.
The story made few waves, but I could never let go of the Vert concept. The more marketing became automated, the more ads learned about us, the closer the world of the Verts seemed. They were already out there in the embryonic form of software, lacking only human form with which to make that crucial personal connection.
Finally, at the Consumer Electronics Show in Las Vegas — appropriately, the fakest city on the planet, filled with its most devilishly effective advertising — I came face to face with what I instantly recognized as Verts. They weren't called that, of course. Star Labs, the Samsung-backed company that launched its human-skinned creations to great fanfare Tuesday, prosaically calls them Artificial Humans.
But each one had a name, and a personality, and an ethnicity, and a job they were demonstrating on a human-sized HD screen at eye level, and they seemed to hold your gaze for just the right amount of time. They were able to answer simple questions. And I had the creepy back-of-the-neck sensation: they felt real. I knew these creatures, and they would conquer the world one day. I knew these creatures, and they would conquer the world one day.
For example, while chatting with a couple of folks from CNN, we realized we were standing in front of an Artificial Human designed to be a journalist, reporting soundlessly into a branded mic. Stick her on local news in front of extreme weather, reading online updates, and she'd pass a Turing test, we realized. As if in response, AI journo walked off screen and came back zipped up in a hooded windbreaker. Woah.
I'm not making any great claims for Neon itself as the bringer of the conquest. The live demos of their capabilities were stilted; the performances on their individual screens were essentially prerecorded bits, though they could respond to some basic voice commands. (For example, they had a special pose when you asked to take a selfie).
"Neon is the equivalent of those early e-ink companies at CES back in the day," as an appropriately skeptical ZDNet writer puts it. "The best demo doesn't exactly win in the long run. Remember, Microsoft had a slate PC years before Apple's iPad."
Yep, they did, and I was here at CES to test it, and I had a firm conviction that the future would look something like that too. It took less than 10 years for Steve Jobs to surpass Microsoft's concept, miniaturizing it into an iPhone and changing the world along the way.
Which company is going to make the iPad to Neon's slate PC? If I knew that, I'd play the stock market in my spare time instead of writing speculative fiction. Neon is at least being smart about keeping its platform open-source and inviting others to contribute; the idea is that any real-life human could be turned into a Vert (sorry, I mean Artificial Human). Neon plans to release an official beta version of this very early software interface at a stage-managed event later this year called Neonworld. It also consciously invokes Jobs' memory with an artificial entrepreneur dressed in turtleneck and jeans.
Point is, the product is inevitable because the economic factors behind it are inevitable. Every organization wants to employ fewer actual humans; every organization wants the magic of human connection with its customers. What if you could do both at the same time? Well, the combination of constantly self-upgrading AI systems and human skin suits like this — which have crossed the uncanny valley because they're stitched together from the performances of actual human actors — will take us there.
It was the range of occupations on display that suggested Neon has seen future needs accurately. There was an airline stewardess you could imagine dealing with a line of pissed-off passengers at the gate. A doctor to ask all the basic questions in your annual check-up. A chef to make you think the restaurant actually cares about your order. A yoga teacher to not judge you. A park ranger to answer questions at a visitor's center — good grief, nobody show this tech to the Department of the Interior. A music teacher and a grade school teacher suggested we could be looking at the future of low-cost public education.
And then there was Monica, a shaven-headed woman who looked younger and dressed hipper than her fellow AI humans. I couldn't figure out Monica's role in all this until Neon demonstrated how good she was at reading a random children's story they fed her. Oh god, I realized with a sinking sensation, Monica's the babysitter of choice for really bad parents.
There were no Neon Artificial Humans working in retail, perhaps because it's too on the nose. But that seems one of the most obvious applications of this technology. It doesn't take much for AI to be more knowledgable, more engaged and less irritating than the average big box store sales clerk (sorry, big box store sale clerks); enough screens and they can even walk you over to the right aisle.
Self-checkout stands at the supermarket are still met with consumer skepticism, but what if there were a friendly human-like being on a full human-sized screen to talk you through the process?
Well, then we'd be one giant leap closer to the world of the Verts: a mundane, Las Vegas-like dystopia of self-aware AI salespeople whose powers of connection have grown greater than the companies supposedly controlling them.
Topics Artificial Intelligence CES
