Ubisoft deployed an AI moderation tool called ToxBuster that reads players' in-game text chat in real time and flags or removes abusive messages across select online games including Rainbow Six Siege. Unlike word-blocklist systems, ToxBuster reads the context of a full conversation before deciding whether a message is harmful. Voice chat moderation launched in Rainbow Six Siege X in October 2025.

ToxBuster was developed by Ubisoft's La Forge research group and uses a BERT-based large language model architecture to analyze chat messages in context, including prior lines of a conversation, rather than evaluating each message in isolation. This allows the system to distinguish between gaming-specific phrases used as strategy and identical phrases used as harmful comments. The system was tested on data from Rainbow Six Siege and For Honor. After ToxBuster-powered automated text chat moderation launched in Rainbow Six Siege Year 9 Season 4 (December 2024), Ubisoft reported a 33% drop in players reported for abusive text chat and an 80% reduction in highly toxic chat volume. Voice chat moderation launched with the Y10S3.3 patch in October 2025, with a community feedback layer allowing players with 'Respectable' or higher reputation standings to confirm or challenge automated detections. Ubisoft also partnered with Riot Games on a research project called Zero Harm in Comms to develop privacy-protecting frameworks for sharing moderation training data across games.