Electronic Arts is looking into a new gameplay option where players can issue voice commands to control otherwise non-controllable or non-playable characters for various actions.
According to a patent filing from last month, EA wants to expand on the idea of controlling characters through just hand gestures or body motion which can be captured by motion-sensing hardware. The concept of selecting from a list of voice commands is not new but which are typically already defined by the developer. What the publishing giant is envisioning is for players to control characters on the fly through actual real-life voice commands.
Through the magic of machine-learning systems, EA wants to create a system where a game can learn in-game character actions through player-defined speeches. By collecting data from multiple players, such a game can hence potentially identify speech and have characters perform in-game actions accordingly.
The filing, for example, points out that saying “duck” or “jump” can have non-playable characters or companions perform the said actions. In similar fashion, a voice command can be issued to have non-playable characters do other actions such as flanking enemy posts. All of these actions however will have to be first learned by the game through player-defined context.
During execution of an instance of a video game, current game state information may be provided to the one or more trained machine learning models based on an indication that a given utterance was spoken by the player. A system may then cause one or more in-game actions to be performed by a non-player character in the instance of the video game based on output of the one or more machine learning models.
While players (or at least veteran players) can easily control multiple characters in a game like FIFA 21 during a sprint, the idea of shouting out commands to have other characters move up into position before shooting can be well imagined.