Most people don't know how a computer does anything, let alone the kernel.

Server-side anti-cheats, like Polar for Minecraft.

The alternative is to create a solution that checks to ensure players aren't doing something that they aren't supposed to do. Sometimes this is easy, like ensuring they don't move too far in a single server tick, or their velocity doesn't get higher or change faster than it should be. Sometimes this is more difficult, like not transmitting the location of enemy players unless they're actually visible to that player. No matter what, it's custom.

Other AC solutions are mostly plug-and-play. They still require some effort, but not nearly the same amount. It's much more appealing to a studio to spend time on developing the actual game, and pay for an AC solution, than to hire people to just handle AC.

There's one more example, that's even more expensive, of using AI detection, which I think Valve still does a lot of. They've been adding this for much longer than the current AI movement has been happening. It takes a lot of data and tagging that data for cheating or not. It's not a perfect solution (Valve does other solutions in addition to this), but it can work really well.

Heuristic data analysis

Its expensive to monitor and store data. Of course you could check and monitor account activity server side whenecer there is a report but youd have to store it all and manually review it.

People do in fact, still manage to cheat on consoles. Also, a locked down ecosystem for games is very far from a solution and comes with many of it's own drawbacks.

Right so the solution is that people should own multiple expensive devices. That seems discriminatory to me

I think they were viewing it from a risk justification perspective. Giving anything kernel level access is high risk, and game publishers have not even remotely earned that level of trust.

Video games are not that important. That's why cheating in them is not that serious. It is not worth the security and privacy risks of putting rootkits in people's machines to address cheating. Many people in gaming communities advocate physical and sexual violence towards people who cheat, there is a perception that gaming is more important than it actually is.

Games aren't important enough to ever outweigh the risks, just like they aren't actually important enough to justify physical and sexual violence against those who break their rules. People trying to argue for this as some paramount issue that needs to be combated no matter the cost need to get some perspective.

Anti-cheat that isn't kernel-level has been around for ages - and even with the kernel level anti-cheat people still manage to cheat. I don't think a special OS with it's own kernel, just for gaming, would be practical. Most people aren't going to install something like that, hell, most people that use a PC have no idea how installing an OS functions. Just have anti-cheat that isn't kernel level, and have a robust report/banning system. Team Fortress 2 was practically unplayable for a long time due to it's botting problems, but they managed to resolve it without kernel level anti-cheat.