A new A.I. system from Facebook’s artificial intelligence research lab has learned how to lie while haggling to get what it wants, according to a report.
“Facebook researchers used a game to help the bot learn how to haggle over books, hats, and basketballs. Each object had a point value, and they needed to be split between each bot negotiator via text,” explained Quartz. “From the human conversations (gathered via Amazon Mechanical Turk), and testing its skills against itself, the A.I. system didn’t only learn how to state its demands, but negotiation tactics as well—specifically, lying.”
“Instead of outright saying what it wanted, sometimes the A.I. would feign interest in a worthless object, only to later concede it for something that it really wanted,” they continued. “Facebook isn’t sure whether it learned from the human hagglers or whether it stumbled upon the trick accidentally, but either way when the tactic worked, it was rewarded.”
“The ability to think ahead is crucial and with the introduction of something called dialog rollouts, which simulate future conversations, the bots were able to do so,” claimed CNET. “Or as FAIR scientist Mike Lewis put it: ‘If I say this, you might say that, and then I’ll say something else.’ Lewis said those rollouts are the key innovation in this project.”
Facebook’s Chief of Artificial Intelligence, Yann LeCun, previously worked in developing advanced surveillance systems.
“LeCun, the Facebook tech guru, was one of four authors of a study aimed at teaching computers how to identify people as they walked down a street, a technique known as pedestrian detection,” wrote The Sun reporter Jasper Hamill in a report last year. “This technique is useful in surveillance and security because it allows computers to quickly analyse CCTV footage and work out if it can see a human or not.”
“LeCun also worked on a study exploring facial recognition, an area of technology focused on allowing computers to detect people’s face. Sophisticated version of this tech can reliably work out someone’s identity,” he continued, before pointing out that Facebook has access to millions of personal photographs of its users.
On Facebook’s Artificial Intelligence Research site, the team claims to “seek to understand and develop systems with human level intelligence by advancing the longer-term academic problems surrounding A.I.”
“Our research covers the full spectrum of topics related to A.I., and to deriving knowledge from data: theory, algorithms, applications, software infrastructure and hardware infrastructure,” they explain. “Long-term objectives of understanding intelligence and building intelligent machines are bold and ambitious, and we know that making significant progress towards AI can’t be done in isolation. That’s why we actively engage with the research community through publications, open source software, participation in technical conferences and workshops, and collaborations with colleagues in academia.”
Charlie Nash is a reporter for Breitbart Tech. You can follow him on Twitter @MrNashington or like his page at Facebook.