Artificial Intelligence and 'Westworld': The Future of Robotics at Silicon Valley Comic-Con

Sunday, 23 April 2017 - 6:28PM
Robotics
Artificial Intelligence
Sunday, 23 April 2017 - 6:28PM
Artificial Intelligence and 'Westworld': The Future of Robotics at Silicon Valley Comic-Con
Image credit: HBO
Everyone who's watched Westworld is looking forward to a future that looks like this:



When in fact, it may look more like this:



Today, we sat in on a panel titled "Artificial Intelligence: Is Westworld a Preview of Our Future?" with J.J. Duncan (pop culture writer), Omar Abdelwahed (Head of Studio for Southbank Robotics), Lonnie Conway (Editor at Zimbio), and two Stanford AI+ Club members, Tim and Bill. The takeaways we got were these:

1. The robots of the future won't look like realistic humans, because it's a lot easier to make personable, empathetic robots like Pepper than uncanny, semi-photoreal humans.
2. It's not the robots we should be worrying about—it's the humans.

Right off the bat, Omar said "I have a difficult time relating to squares or cylinders, or voices in the air that don't attach to anything." He was referencing the Amazon Echo and Microsoft's Cortana, which are essentially personal assistants built into devices. He pointed to Southbank's flagship robot, Pepper, as a better alternative: between her admittedly Disney-esque giant eyes, child-sized stature, and human shape, she feels like an actual presence in a room, similar to a person. Omar pointed out that she's always doing something, rather than sitting completely motionless unless called on, which makes her even more human-like. Another touch Southbank has given Pepper is the primary way to turn her off: instead of saying "Pepper turn off," you place your hand on top of the robot's head for a few seconds, until it sighs and goes to sleep.

All of this creates a sense of empathy for Pepper, but it's not Southbank's goal to make a robot that's indistinguishable from a human—first of all, there's not enough use for that yet. The park in Westworld aimed to create a lifelike, consequence-free environment for thrill-seekers and rich hedonists, but that's a very specific goal. Right now, most people are fine with being able to tell robots apart from humans and acting accordingly.

Instead of human verisimilitude, Omar says, we should aim to create characters (fittingly enough, Pepper has a second, manual off-switch at the base of the neck because that's where Data's off-switch was in Star Trek). Still, some semblance of humanity is needed to generate empathy, and one of the features Pepper possesses is ominously called "autonomous life," which allows her to dynamically change her gestures and speech patterns, so that no interaction is ever the same.



According to Omar, Japan already has over 7,000+ Pepper robots in homes and small businesses, meant to act as 'social robots'—they'll greet you when you come home, remind you of upcoming events, and carry on conversations. Pepper is part of the Internet of Things, meaning that it can check to see if your friends are at a concert nearby and suggest joining them, or let you know when you need to get more milk.

So what do we do with all these programmable robot companions? What happens when we're able to upgrade them to something that resembles a human mind? This is where Omar warns "What you should worry about is not AI, what you should worry about is people, and what they do with AI." A topic brought up during the panel was Microsoft's Tay Twitter bot, which quickly went from a friendly chatterbot to a racist icon based on its interaction with trolls. There is always the danger of AI changing itself faster than we can control it, but in the meantime, the real issue is keeping humans from abusing the technology for human ends. Westworld sums it up best: violent delights have violent ends. 
 
Science
Science of Sci-Fi
Robotics
Artificial Intelligence