Facebook has been working to improve Artificial Intelligence or AI and in the process has made some pretty significant accomplishments. While Facebook has been working with bots, teaching them negotiation skills in order to replace human negotiations and so far, experiments have proved to be largely successful. However, as one of the negative effects of the research, the bots have instead designed their own language and continue conversations in the new artificial language. Also Read - Facebook Gaming launches Pac-Man Community: Features online multi-player modeAlso Read - Big threat for Google, Twitter, WhatsApp, and LinkedIn users
The research was published in detail in a Facebook blog post where the social media giant talks in detail regarding the future of bots and how they can replace a lot of human element when it comes to mundane tasks. The bots were taught machine learning and then, post the learning, allowed to converse freely with each other to create their own negotiation skills. However, in the process, they ended up creating a completely different language set, one that is not a part of any known human language and that cannot be understood by the layman. Also Read - Facebook Messenger introduces a new Split Payments feature: How to use it
Facebook conducted in-depth research when it comes to the bots and the negotiation behavior left the researchers surprised. The bots proved to be extremely smart and efficient in picking up on human behavior and even copied it multiple times. The bots would often pretend to be interested in something and then, accept to let go as a “compromise” in order to get to what they really wanted in the first place. In negotiating with humans, many times bots easily surpassed their craftiness and were able to successfully complete the bargain. ALSO READ: After Facebook, YouTube reveals how it is using AI to curb terrorism
There were multiple situations and tests that Facebook put the bots through and the bots learned to anticipate a response. The bots would think ahead and were able to arrive at conclusions much before, and thus, change the direction of the conversation and the negotiation accordingly. Even longer conversations proved to be no hassle for the bots and they were able to successfully complete the entire situation. There were three main areas in which the bots focused: Negotiating harder, where bots continued the conversation till the desired result was achieved; Intelligent maneuvers, where they chose tact over aggression to deal with situations and; Producing novel sentences, where bots did not sound repetitive but were able to come up with innovative sentences.
Facebook explains in its blog post, “To date, existing work on chatbots has led to systems that can hold short conversations and perform simple tasks such as booking a restaurant. But building machines that can hold meaningful conversations with people is challenging because it requires a bot to combine its understanding of the conversation with its knowledge of the world, and then produce a new sentence that helps it achieve its goals.”
“Today, researchers at Facebook Artificial Intelligence Research (FAIR) have open-sourced code and published research introducing dialog agents with a new capability the ability to negotiate. Similar to how people have differing goals, run into conflicts, and then negotiate to come to an agreed-upon compromise, the researchers have shown that it s possible for dialog agents with differing goals (implemented as end-to-end-trained neural networks) to engage in start-to-finish negotiations with other bots or people while arriving at common decisions or outcomes,” it added.