# US 2026/0097313 A1 · Cooperative game artificial intelligence (ai) voice chat

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0097313 A1
- Document ID: US20260097313A1
- Status: Published application
- Kind: A1
- Application number: 18/908,704
- Filed: 2024-10-07
- Published: 2026-04-09
- Earliest claimed: 2024-10-07
- Assignees: Sony Interactive Entertainment Inc.
- Applicants: Sony Interactive Entertainment LLC
- Inventors: Sean Whitcomb
- Web page: https://patentgrep.com/patent/US20260097313A1

## CPC

- A63F 13/67: adaptively or by learning from player actions, e.g. skill level adjustment or by storing successful combat sequences for re-use
- A63F 13/54: involving acoustic signals, e.g. for simulating revolutions per minute [RPM] dependent engine sounds in a driving game or reverberation against a virtual wall
- A63F 13/87: Communicating with other players during game play, e.g. by e-mail or chat
- A63F: CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR

## Abstract

Methods and systems are provided for generating chat communication in a video game. One method includes accessing a gameplay database that includes dialog scripts from player interactions occurring during interactive scenarios of a cooperative video game. The method includes processing the dialog scripts to extract chat words between real players during the cooperative video game and labeling the chat words to correlate the chat words to one or more interactive scenarios and correlate the chat words to a game context, wherein over time an artificial intelligence (AI) model is trained using the chat words that are labeled to learn contextually relevant chat words for the cooperative video game. The method includes identifying a non-player character (NPC), during an instance of gameplay of the cooperative video game, the NPC is present in the cooperative video game while an avatar of a real player is present in a scene of the cooperative video game. The method includes generating one or more NPC chat words for the NPC using the AI model, the NPC chat words being audible to the real player.

## Claims

### Claim 1

A method for generating chat communication in a video game, comprising,
accessing a gameplay database that includes dialog scripts from player interactions occurring during interactive scenarios of a cooperative video game;
processing the dialog scripts to extract chat words between real players during the cooperative video game and labeling the chat words to correlate the chat words to one or more interactive scenarios and correlate the chat words to a game context, wherein over time an artificial intelligence (AI) model is trained using the chat words that are labeled to learn a chat style for the cooperative video game;
identifying a non-player character (NPC), during an instance of gameplay of the cooperative video game, the NPC is surfaced in the cooperative video game while an avatar of a real player is present in a scene of the cooperative video game;
detecting a location of the NPC in the scene to determine if the NPC is within a chat proximity to the real player; and
generating one or more NPC chat words for the NPC consistent with the chat style of the cooperative video game using the AI model, the NPC chat words being audible to the real player and visually depicted as being output by the NPC toward the real player.

### Claim 2 (depends on 1)

The method of claim 1, wherein the labeling of said chat words occurs after a normalizing function to filter out words from the dialog scripts, wherein the normalizing function assists in identifying a relevance factor of the words in relation to the interactive scenarios and in relation to the game context, such that one or more of the words are not used as input to the AI model for training.

### Claim 3 (depends on 2)

The method of claim 2, wherein the labeling includes tagging single words or groups of words from the chat words with context markers that assist in relating the chat words to said interactive scenarios and relating the chat words to the game context.

### Claim 4 (depends on 3)

The method of claim 3, wherein said training of the AI model enables the AI model to predict chat words that are appropriate for a current interactive scenario based on real-time context analysis,
wherein the real-time context analysis identifies a time window during which said NPC is enabled to output said NPC chat words, wherein if the time window is expired for said current interactive scenario, then said NPC will not output said NPC chat words even when the real player is within the chat proximity to the NPC.

### Claim 5 (depends on 1)

The method of claim 1, wherein training said AI model further includes providing the AI model with custom chat words that relate to the cooperative video game or specific cooperative video games to fine tune the AI model.

### Claim 6 (depends on 1)

The method of claim 1, wherein the identifying the NPC in the scene of the cooperative video game includes game state monitoring, wherein the game state monitoring includes analysis of game state during interactivity of the cooperative video game.

### Claim 7 (depends on 1)

The method of claim 1, wherein the avatar of the real player is controlled by the real player in the scene and other scenes of the cooperative video game and the avatar is made to interact in one or more of the interactive scenarios, where said interaction in the one or more of the interactive scenarios by the avatar continually updates said game context.

### Claim 8 (depends on 1)

The method of claim 1, wherein during said interaction by the avatar in the one or more interactive scenarios data is processed by the AI model to identify said NPC chat words.

### Claim 9 (depends on 8)

The method of claim 8, wherein real-time context analysis identifies a time window during which said NPC is enabled to output said NPC chat words.

### Claim 10 (depends on 8)

The method of claim 8, wherein if the time window is expired, then said NPC will not output said NPC chat words toward the avatar of the real player even when the real player is within the chat proximity to the NPC.

### Claim 11 (depends on 1)

The method of claim 1, wherein the NPC chat words are throttled in output frequency based on an engagement level of the real player that controls the avatar, the engagement level is in part based on actions currently being taken by the player in the scene and state data analysis that identifies actions being taken by the player or required to be taken by the player in the interactive scenarios of the cooperative video game.

### Claim 12 (depends on 1)

The method of claim 1, wherein the NPC chat words are throttled based on a user preference setting that identify a level of engagement desired for the NPC or other NPCs in the scene during the interactive scenarios of the cooperative video game.

### Claim 13 (depends on 12)

The method of claim 12, wherein the user preference setting includes an interactive sliding scale that sets the level of engagement desired for the NPC or other NPCs, wherein the level of engagement indicates whether the NPC or other NPCs should more aggressively or less aggressively chat with the avatar of the real player.

### Claim 14 (depends on 1)

The method of claim 1, wherein the AI model is used to modify the NPC chat words for urgency triggered by the game context, wherein more urgency requires the NPC chat words to be delivered by the NPC to the real player in a language format that is briefer and more specific.

### Claim 15

A method for generating chat communication in a video game, comprising,
accessing a gameplay database that includes dialog scripts from player interactions occurring during interactive scenarios of a cooperative video game;
processing the dialog scripts to extract chat words between real players during the cooperative video game and labeling the chat words to correlate the chat words to one or more interactive scenarios and correlate the chat words to a game context, wherein over time an artificial intelligence (AI) model is trained using the chat words that are labeled to learn contextually relevant chat words for the cooperative video game;
identifying a non-player character (NPC), during an instance of gameplay of the cooperative video game, the NPC is present in the cooperative video game while an avatar of a real player is present in a scene of the cooperative video game;
generating one or more NPC chat words for the NPC using the AI model, the NPC chat words being audible to the real player.

### Claim 16 (depends on 15)

The method of claim 15, wherein a location of the NPC in the scene is identified to determine if the NPC is within a chat proximity to the real player.

### Claim 17 (depends on 15)

The method of claim 15, wherein said training of the AI model enables the AI model to predict chat words that are appropriate for a current interactive scenario, and a time window is identified during which said NPC is enabled to output said NPC chat words.

### Claim 18 (depends on 17)

The method of claim 17, wherein if the time window is expired for said current interactive scenario, then said NPC will not output said NPC chat words even when the real player is within the chat proximity to the NPC.

### Claim 19 (depends on 15)

The method of claim 15, wherein the NPC chat words are throttled in output frequency based on an engagement level of the real player that controls the avatar, the engagement level is in part based on actions currently being taken by the player in the scene and state data analysis that identifies actions being taken by the player or required to be taken by the player in the interactive scenarios of the cooperative video game.

### Claim 20 (depends on 15)

The method of claim 15, wherein the NPC chat words are throttled based on a user preference setting that identify a level of engagement desired for the NPC or other NPCs in the scene during the interactive scenarios of the cooperative video game.
