GDC 2025 Session Viewer

This talk explores the methodologies and challenges of implementing server side Machine Learning (ML) based aimbot detection and its impact on online game security.
The rise of competitive online gaming, particularly in Call of Duty has led to an increase in cheating, with aimbots being one of the most disruptive forms. Aimbots automatically lock onto targets, giving players an unfair advantage and eroding trust within the gaming community.
Traditional detection methods, like heuristic systems, struggle to keep up with increasingly sophisticated aimbots.
ML offers a robust solution by analyzing gameplay data to recognize patterns indicative of aimbot use. By training models on labelled datasets and focusing on angle velocity, acceleration etc. Activision researchers achieved over 85% detection accuracy while minimizing false positives.
Challenges remain, such as collecting more representative data and avoiding misclassifying high-skilled players, but ML is proving to be a powerful tool in maintaining fair play in competitive gaming.

Takeaway

Attendees would gain an understanding of various types of aimbots used in FPS games.
Attendees would also learn about the application of advanced machine learning techniques used by Call of Duty security research team to detect these aimbots and the impact of this detection on the Call of Duty franchise.

Intended Audience

Game developers, Security researchers, First person shooter game enthusiasts, ML enthusiast