'We're Nowhere Near Skynet Today,' Says Top Facebook Engineer

Great news, everyone! Facebook isn’t on the verge of bringing about an apocalyptic future wherein robots rise up and enslave humanity. At least not according to one of Facebook’s chief engineers.

At MIT Technology Review’s Digital Summit Tuesday morning, editor in chief Jason Pontin asked Jay Parikh, Facebook’s vice president of infrastructure engineering, whether the rest of us ought to be just a little nervous that Facebook is developing artificial intelligence that uses “deep learning” techniques that mimic the processes of the human brain. AI that gets smarter on its own and achieves self-awareness is a key plot point of dark-future science fiction movies like “The Terminator” and The Matrix,” along with worldwide networks of autonomous machines.

Still quite a long way off. (Photo credit: Wikipedia)

(Did I mention that Facebook is also moving ahead with plans to blanket the world with a web of high-flying drones? Or that Google just bought a company that makes satellites, Skybox, for $500 million?)

Parikh responded by saying it would take an awfully long dotted line to connect Facebook’s current work on AI, which is mostly around improving search and recognition of images, particularly faces, to something like Skynet, the digital entity that rebels against humanity in “The Terminator.”

“The team thinks a lot about the ethical issues that crop up here,” Parikh said. “We’re nowhere near Skynet today. That’s way the heck out there. That’s not close to anything that’s possible today.”

So we’re safe for now — unless, of course, the singularity of the distant future devises a way to send the killer robots back in time.