For the longest time, robots have had one vital weakness – they couldn’t lie. You could ask a robot if it was going to kill you and it would have to truthfully answer. It would still kill you, but at least you knew it was coming. Thanks to some researchers at Georgia Tech, robots may soon be able to trick you into thinking everything is fine until it fires a laser in your face.
With funding from the Office of Naval Research, Professor Ronald Arkin and others have created robots that can lie to each other. The research came to this breakthrough after observing how animals in the wild deceive others to protect what’s important to them. In this case, the researchers used squirrels that intentionally patrol empty patches of land to deceive other squirrels into thinking it has hidden acorns there.
The first successful test of the robots had them performing the exact same deception strategy. One robot lured the other towards false locations to protect resources. Professor Arkin believes this could lead to better protection of vital military resources.
“This application could be used by robots guarding ammunition or supplies on the battlefield,” said Arkin, a Regents Professor in Georgia Tech’s School of Interactive Computing. “If an enemy were present, the robot could change its patrolling strategies to deceive humans or another intelligent machine, buying time until reinforcements are able to arrive.”
For now, the robots are nothing more than little vehicles with laptops strapped to them. It has a long way to go before they start deceiving humans to walk into a secluded “cave of treasures” that actually turns out to be a death trap. Actually, you probably should never walk into a “cave of treasures” even if the robot seems like a nice guy.
Even if the robots can lie, humanity can still gain the upper hand with an EMP grenade or other related weaponry. Professor Arkin must have thought about that scenario as he came up with another method of deceit that would allow a robot to look tougher than it actually is. He used the Arabian Babbler as an example that joins other birds in harassing a predator to scare it away despite being far weaker than the predator in question.
“In military operations, a robot that is threatened might feign the ability to combat adversaries without actually being able to effectively protect itself,” said Arkin. “Being honest about the robot’s abilities risks capture or destruction. Deception, if used at the right time in the right way, could possibly eliminate or minimize the threat.”
Great, so now we have robots that can deceive humans while making themselves look even more threatening than they really are. This Professor Arkin must really have it in for humanity. That’s actually not the case as he himself calls for discussion over the use of deception in non-military machines. The only caveat here is that he is still the guy who’s giving military robots, the robots that have weapons mind you, the ability to lie.
[h/t: Wired UK]