Recent updates to the Steam client have revealed references to a potential AI system named 'SteamGPT', which Valve may be developing to bolster security on its platform. The leaked files suggest that this system could automate the evaluation of in-game incidents and analyze suspicious account activities. Specifically, the files mention terms associated with generative AI, such as multi-category inference and fine-tuning, indicating a sophisticated approach to handling incident reports in multiplayer games. Functions labeled as 'SteamGPTSummary' are designed to summarize suspicious account histories, taking into account factors like VAC bans, Steam Guard, and account lockdowns. This AI could also evaluate evidence from user accounts, such as email addresses and the origin of linked phone numbers, to identify fraudulent behavior. Additionally, the system may utilize an account's trust score to enhance matchmaking security in games like Counter-Strike 2. Overall, 'SteamGPT' represents a significant step towards integrating AI into gaming security, potentially improving user safety and experience on the platform.