AI Agent
Last updated
Last updated
We designed rich AI behavior trees that fit for various scenarios, and arranged distinctive AI behaviors for different types of players. AI players are able to simulate all the behaviors of actual players, such as jumping on hidden spots and sneaking up on real players. Therefore, it is hard to even distinguish them from actual players.
When teaming up with actual players, AI players can cooperate with their teammates, receiving information from and interacting with them. AI players have a super diverse tactical system that can make real-time responses according to the battle.
Based on AI players, we provide players with a "Human vs AI player" mode for training and learning, which helps players better understand the game rules.
In the future, users' game data can be stored on the chain, and they can use different LLMs to train their own AI players, or they can rent their own data for other players to train. User's AI players can also participate in related competitions in the future and receive rewards to make money for themselves.
We will use AI learn from cheating behavior data, to improve our Anti-cheating Inspector System to detect players who use cheating programs more efficiently.
