Skip to Content
Artificial intelligence

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

It could be a step toward building AIs that can handle complex problems requiring compromise.

""
DALL.E 2 using the prompt “AI conquering Europe in a boardgame.”Melissa Heikkilä via DALL.E 2

Meta has created an AI that can beat humans at an online version of Diplomacy, a popular strategy game in which seven players compete for control of Europe by moving pieces around on a map. Unlike other board games that AI has mastered, such as chess and Go, Diplomacy requires players to talk to each other—forming alliances, negotiating tactics—and spot when others are bluffing. 

The AI, called Cicero, ranked in the top 10% across 40 online games against 82 human players (who were not aware they were competing against a bot). In one eight-game tournament involving 21 players, Cicero came first. Meta described its work in a paper published in Science

Learning to play Diplomacy is a big deal for several reasons. Not only does it involve multiple players, who make moves at the same time, but each turn is preceded by a brief negotiation in which players chat in pairs in an attempt to form alliances or gang up on rivals. After this round of negotiation, players then decide what pieces to move—and whether to honor or renege on a deal.

At each point in the game, Cicero models how the other players are likely to act based on the state of the board and its previous conversations with them. It then figures out how players can work together for mutual benefit and generates messages designed to achieve those aims.

To build Cicero, Meta marries two different types of AI: a reinforcement learning model that figures out what moves to make, and a large language model that negotiates with other players.

Cicero isn’t perfect. It still sent messages that contained errors, sometimes contradicting its own plans or making strategic blunders. But Meta claims that humans often chose to collaborate with it over other players. 

And it’s significant because while games like chess or Go end with a winner and a loser, real-world problems typically do not have such straightforward resolutions. Finding trade-offs and workarounds is often more valuable than winning. Meta claims that Cicero is a step toward AI that can help with a range of complex problems requiring compromise, from planning routes around busy traffic to negotiating contracts.

Deep Dive

Artificial intelligence

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

Google DeepMind’s new generative model makes Super Mario–like games from scratch

Genie learns how to control games by watching hours and hours of video. It could help train next-gen robots too.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.