It's a cat-and-mouse game. A detector detects inhuman speed? Aimbots adapt by limiting movement speed and reaction time. Detectors start looking for unnaturally smooth movement? Aimbots add randomness to make the movements less consistent and smooth.
Valve was talking about using machine learning to recognize cheating back in 2017: https://www.pcgamer.com/valve-wants-to-take-a-machine-learni... and more info from 2018: https://www.youtube.com/watch?v=ObhK8lUfIlc
And this one from 2023 is about using an RNN Deep Learning model for cheat detection: https://www.usenix.org/conference/usenixsecurity23/presentat...
It's a cat-and-mouse game. A detector detects inhuman speed? Aimbots adapt by limiting movement speed and reaction time. Detectors start looking for unnaturally smooth movement? Aimbots add randomness to make the movements less consistent and smooth.