The push back on kernel level anti-cheat on security grounds has always felt odd to me. If you don't trust them to run kernel level code why do you trust them to run usermode code as your user? A rogue anticheat software could still do enormous damage in usermode, running as your user, no kernel access required.

Being in kernel mode does give the rogue software more power, but the threat model is all wrong. If you're against kernel anti-cheat you should be against all anti-cheat. At the end of the day you have to chose to trust the software author no matter where the code runs.