When playing online games, you can always encounter cheaters instead of skilled players. The Anti-Cheat Police Department (@AntiCheatPD) shared a gameplay video of Call of Duty demonstrating a new cheating method using machine learning. This system utilizes a PC and a capture card to store gameplay footage for the algorithm to learn from.
Players still need to play the game to control their characters, but the algorithm adjusts itself when aiming to ensure hitting the target. AntiCheatPD describes the algorithm as similar to “aim assist,” but it is enhanced significantly, requiring the player to do nothing but aim randomly, and the machine will shoot for them.
The headache for developers is that the system will struggle to detect cheating because there are real players controlling the characters, but their reflexes are on par with the best players. So how can we tell who is cheating and who is truly skilled? Only time will tell if this system has any weaknesses that can help us detect it.
It is unclear how input lag affects this new cheating method. Clearly, the gameplay video must go through the capture card before being processed by the algorithm, then it sends signals to the controller input, so there might be a few milliseconds delay. But if it works well enough, console games will likely not escape this threat. You can easily connect a console to a computer via a capture card, so this risk will significantly affect all platforms.
According to PCMag