This post is based on the Intelligence Rising Game I played with members of THEA, PISE, EA Delft & others. I found the game very entertaining thanks to the players and game makers and I hope others will try the game out and share their experiences here. Enjoy!
The game is for 8-20 players and 1-3 gamemakers. The players and game makers should have a basic understanding of AI development and the risks of AGI.
The players are separated into 4 groups, representing
- the US Government,
- the Chinese Government,
- Tencent Holdings Ltd.,
- Alphabet Inc.,
- and the AI Alignment Community.
Each group gets a set of Goals they strive to achieve throughout the game, Resources which they can use to achieve said goals and Weaknesses that limit their actions. These are made by the game makers, should resemble real life and must be kept secret from the other groups.
The game consists of 5 rounds, each symbolizing a 3-year period starting from the present. At the beginning of each round, the game makers present 5-8 scenarios based on the players' actions in the previous round(s). The first round starts with scenarios that already happened or will likely happen in a few months. Every round consists of
- 20 minutes of parallel arrangements within and between groups and the communication of up to one private action per group only to the game makers,
- 10 minutes for presenting up to two public actions per group to everyone,
- (if needed) an additional 5-10 minutes for the game makers to develop new scenarios for the next round, based on the public and private actions and additional imaginary events.
The arrangements and actions should align with the Goals, Resources and Weaknesses of each group.
The game ends after the 5th round and the game makers conclude by one last set of scenarios (outcomes of actions). Optional elements:
- declaring a winner(s) based on which group could achieve the Goals set at the beginning,
- sharing the secret actions or other behind-the-scenes elements with the players.
