The perception of cheating levels is orders of magnitude more than the actual amount. The vast majority of games and players are clean.
https://support.chess.com/article/648-what-do-i-need-to-know-about-fair-play-on-chess-com
https://www.chess.com/article/view/online-chess-cheating
https://www.chess.com/article/view/chesscom-update-june-2025#FairPlay
The site closed more than 113,000 accounts in May for fair play violations and has a full team of staff working on reports along with some automated systems for detection.
https://www.chess.com/article/view/fair-play
That said, discussions of cheating, potential cheating, or cheat detection are not allowed in the general forums. If you would like to discuss join the following club
https://www.chess.com/club/cheating-forum
Even though Chess.com has the best anti-cheat engine in the world, there are still a lot of cheaters on the site. If every player plays honestly, what will happen to our rating?
First, we assume that cheating players can make 30% of their games undefeated, unless they encounter other cheating players. In the other 70% of the games, they will mix the AI and their own moves, and then they can be defeated. We call their strength at this time mixed ELO. The Elo in Chess.com (called ELOc) and mixed ELO can be exchanged. For example, a cheater with a mixed ELO 2000 would have an ELOc 2159 (ELO 2000 vs ELO 2169, 70 games, win 20 loss 50, another 30 game win with the pure AI).
However, the game between cheaters does not produce a huge ELO difference. We assume A and B play 1000 games. The probability that two cheaters use AI throughout the whole game is 0.3*0.3=0.09. So there is 90 games, and each win 45 games. A use AI for the whole game and B not, there are 0.3*0.7*1000=210 games. Therefore, A will win 210+45=255 games at least. It implies that the maximum diffenence of ELO between cheater is 208. When there are a large number of cheaters, the games between these cheaters will reduce the ELO gap between users.