Have you ever been negotiating with someone and, when you reached an impasse, just started making up a new way of communicating? That’s what the chatty A.I. robots — or “dialog agents” — over at Facebook recently did.
In a story for The Atlantic posted last week, Adrienne LaFrance writes about a remarkable bit of “buried” information found in a recent report by Facebook’s Artificial Intelligence Research Lab. In the report, Deal or No Deal? End-to-End Learning for Negotiation Dialogues, the researchers describe employing machine learning to train their “dialog agents” to negotiate.
Apparently, these A.I. “chatbots” are quite good at making deals, according to the report.
“At one point, the researchers write, they had to tweak one of their models because otherwise the bot-to-bot conversation ‘led to divergence from human language as the agents developed their own language for negotiating,’ ” LaFrance reported. “They had to use what’s called a fixed supervised model instead.
“In other words, the model that allowed two bots to have a conversation — and use machine learning to constantly iterate strategies for that conversation along the way — led to those bots communicating in their own non-human language. If this doesn’t fill you with a sense of wonder and awe about the future of machines and humanity then, I don’t know, go watch Blade Runner or something.”
By Kevin Hyde, Content Writer, Capture Higher Ed