Facebook is building chatbots that think more like humans

Mark ZuckerbergFacebook has developed an AI that is capable of negotiating and developing meaningful conversations, much like a human would, paving the way for more advanced and responsive chatbots.

Researchers at Facebook Artificial Intelligence Research (FAIR) have open-sourced code that introduces the ability to negotiate, giving brands the chance to harness the power of a more human-like chatbot.

“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,” said the FAIR team in blog post.

The AI unit trained the ‘dialog agents’ using a crowdsourced collection of negotiations between pairs of people. These individuals were shown objects and a value for each, and then asked to agree on how to divide the objects between them. Researchers then used their exchanges to teach the AI how to imitate people’s actions when negotiating. The agent was then made to achieve the goals of the negotiation by having the model practice thousands of negotiations against itself.

Bringing this altogether, the agents were then tested in online conversations with people, and the outcome was a positive one.

“Interestingly, in the FAIR experiments, most people did not realise they were talking to a bot rather than another person — showing that the bots had learned to hold fluent conversations in English in this domain,” said the FAIR team. “The performance of FAIR’s best negotiation agent, which makes use of reinforcement learning and dialog rollouts, matched that of human negotiators. It achieved better deals about as often as worse deals, demonstrating that FAIRs bots not only can speak English but also think intelligently about what to say.”