Meta’s game-playing AI can make and break alliances like a human

Learning to play Diplomacy is a big deal for many reasons. Not only does it involve multiple players, who make moves at the same time, but each turn is preceded by a brief bargain, in which players chat in pairs to try to form an alliance or Team up with opponents. After this round of negotiations, the players decide which pieces to move — and whether to honor or abandon the deal.

At each point in the game, Cicero models how other players are likely to act based on the state of the board and the board’s previous conversations with them. It then figures out how players can work together for the common good and creates messages designed to achieve those goals.

To build Cicero, Meta combined two different types of AI: a reinforcement learning model for figuring out what moves to make and a vast language model for negotiating with other players.

Cicero is not perfect. It still sends out flawed messages, sometimes contradicting its own plans or making strategic mistakes. But Meta claims that humans often choose to cooperate with it over other players.

And that’s important because while games like chess or Go end with winners and losers, real-world problems often don’t have such a simple solution. Finding trade-offs and alternatives is often more valuable than winning. Meta claims that Cicero is a step towards AI that can help solve a range of complex problems that require compromise, from planning routes around heavy traffic to negotiating contracts.


News5h: Update the world's latest breaking news online of the day, breaking news, politics, society today, international mainstream news .Updated news 24/7: Entertainment, the World everyday world. Hot news, images, video clips that are updated quickly and reliably

Related Articles

Back to top button