A constant demand that every Counter-Strike player now and in the past has always had is for Valve to improve its Anti-Cheat system. Valve Anti-Cheat (VAC) is the current cheat detection system being used by Valve for all its online competitive games but this system could use a major upgrade and a bunch of computer scientists from the University of Texas at Dallas have come up with the perfect solution, an innovative AI (Artificial Intelligence) powered Anti-Cheat system.
AI Anti-Cheat System For CS:GO
In a bit to detect cheaters who spoil online competitive games for everyone by using third-party software’s to bypass certain restrictions which grant them an unfair advantage, a few computer scientists came together to create an AI anti-cheat system which can work with any MMO (Massive Multiplayer Online) game.
Their own experiment and approach were however based on a study conducted using the popular competitive shooter CS:GO. What the researchers basically did for the study was that they selected 20 students at random and downloaded CS:GO along with three software cheats: AimBot, Speed Hack, and Wallhack.
These players then proceeded to play on a dedicated server set up by the researchers to isolate them from the general online traffic. All the game data to and from the dedicated server was then analyzed by the researchers who noticed certain features such as,
- Number of incoming and outgoing packets of information.
- Size of these packets which depends on the content of information carried within.
- Transmission time.
- The direction of transmission.
- Number of packets in a burst, which is nothing but a group of consecutive packets.
Having analyzed all these features and by constantly monitoring the data traffic the researchers identified certain patterns which indicated cheating. This information was used to train a machine-learning model, a sort of an AI whose aim was to detect cheaters based on these patterns.
Despite the research being conducted on a small set of gamers the researchers stated that their statistical model can be adjusted as per the requirement of the client or the developer. They claimed that gaming companies could even use their own data to train the AI powered anti-cheat and set it up to take appropriate action after detecting a cheater.
Dr. Latifur Khan who is an author of the study said that “Players who cheat send traffic in a different way, we’re trying to capture those characteristics.” Having said that the team of researchers added that they will continue their research and work towards creating an approach that does not require a client-server architecture.
Though the research looks fantastic it is still a bit early to say that this innovative approach to creating an AI-powered anti-cheat is the best step forward as it remains to be tested in a proper gaming environment. It will be interesting to see if any game developer approaches this team of researches to put their anti-cheat to practical use and work in collaboration with them.