In Brief
Cicero, the first AI to achieve human-level competence in the strategy board game Diplomacy
The Trust Project is a worldwide group of news organizations working to establish transparency standards.
Meta, a company that is best known for its work on AI, has announced a new artificial intelligence platform that it says can outperform humans in the realm of diplomatic negotiation. Meta says the system, called Cicero, is the first AI that can reason about, learn from, and intervene in complex social situations.

Diplomacy is more than just a game; the majority of the gameplay is focused on social skills. Natural language is used by players to communicate, negotiate, deceive, form connections, and persuade. A computer player finds it tough to win.
Cicero learned this by playing the online version of Diplomacy at webdiplomacy.net. It eventually mastered the game, scoring “more than double the average” for human players and finishing in the top 10% of players across multiple games, according to reports. AI models for strategic thinking (like AlphaGo) and natural language processing were merged to develop Cicero by Meta (similar to GPT-3)
Cicero makes predictions about other players’ behavior during each game by analyzing the state of the playing field and the previous conversations. It creates a strategy, which it then executes using a language model that can produce dialogue that sounds human, enabling the AI to work with other players.
- Meta suggests that her Cicero research could “weaken communication barriers” between humans and AI, such as holding a long-term conversation to teach someone a new skill.
- It might be the driving force behind a video game where NPCs can converse like people, picking up on the player’s intentions and changing as they go.
- There is a downside to this as well: This technology could be used to influence people by impersonating them and tricking them in potentially harmful ways.
Both the Cicero code and a comprehensive website detailing how it functions have been made available on GitHub by Meta.
Read related articles:
Disclaimer
Any data, text, or other content on this page is provided as general market information and not as investment advice. Past performance is not necessarily an indicator of future results.