When Bots Go Bad

"When Bots go Bad" is a game about the role of Artificial Intelligence in the future of society. The game tasks the player with finding the optimal balance between the usefulness of their AI system and the potential risk of it causing a major catastrophe. The player must explore this idea through two perspectives: that of the engineers developing the machine learning models of the future, and that of the lawmakers and politicians in charge of passing legislation around this rapidly evolving technology. The model architectures featured in the game are based on the real-world architectures used in AI models today, and the laws available for the player to pass are based on suggestions from research papers published by the Global Partnership for Artificial Intelligence (based in Paris) and the Center for AI Safety (based in San Francisco). After the player has developed their AI model and passed the laws that they have chosen, the game fast-forwards to the year 2100 where the player can see how useful their AI was to people and its likelihood of causing a major catastrophe.