> With the way that computer vision and AI continue to improve, I imagine that we will soon have completely external and undetectable cheating peripherals, simply capture the screen direct from the display output, and pass inputs via mimicking a usb human input device.

This already exists. You can stream your screen to another machine running image recognition and pass your mouse input through a controller that injects auxiliary input (there are off-the-shelf products like kmbox, you can make your own as well).

However, it is very important to understand that only a tiny percentage of cheaters in games end up being determined enough to go through hoops to purchase hardware for it (it's much more expensive and not as simple as getting instant gratification by double clicking on an executable). It's basically considered a win to push people into needing go to such lengths to cheat without getting banned.

For hacks that don't access program memory directly, would external hardware make things less detectable? I don't know how anti-cheat programs work but I'd be surprised if they banned every skilled player that happens to be running AHK and OBS. More likely they work with heuristics that try to detect super-human mouse movements, precision/speed-wise.

> For hacks that don't access program memory directly, would external hardware make things less detectable?

Yes. But, it's a game of cat and mouse. Anticheat always is. There's not been a need go to the level of "there are X players who report having a Razer deathadder, which we've validated as having a DPI range of Y-Z. With your OS settings, it's not possible to have achieved that with that mouse therefore we suspect you of cheating". But we're probably there now.

> More likely they work with heuristics that try to detect super-human mouse movements, precision/speed-wise.

It's heurestics all the way down. You're unlikely to be banned for AHK+OBS, but a heurestic on what you're doing with the combo might ban you.

> However, it is very important to understand that only a tiny percentage of cheaters in games end up being determined enough to go through hoops to purchase hardware for it

Today. That will change when it's a cheap off-the-shelf product sold on Amazon. I almost wish we were there already to get rid of this KLA nonsense.

Not to mention, every KLA game I've played has had cheaters regardless. Circumventing the software seems to become a game unto itself for those people.

And all of that for something that is worse - software cheats usually get access to more information than just the pixels rendered to the screen. Seeing through solid walls and stuff like that.

You can do DMA cheating if that’s a non-negotiable for you but it is far more expensive. You install a direct memory access (DMA) card on your “clean” machine and flash it with firmware that hides the fact that it’s a DMA card since some rootkit level anticheats will look for that.

The card then sends memory data to a second computer (dirty) which has the cheat software installed but no rootkit anticheat. The dirty computer can then read all the memory it wants, look through walls, shows enemy health, etc. but on a second computer and monitor. Pretty sure there’s hardware to blend both monitor outputs, though most people do two monitors I believe.

> You can do DMA cheating if that’s a non-negotiable for you but it is far more expensive.

Cheap DMA cards go for couple hundred dollars which isn't all that much. Although unlike what most people tend to believe, the DMA cards do end up getting detected quite often.

> Pretty sure there’s hardware to blend both monitor outputs, though most people do two monitors I believe.

It's called a "fuser".

are there out-of-the-box software tools to do this?

I thought about something like this before, but its non-trivial to train an image recognition system to do this, let alone the commands to mimic it through hardware, its quite an engineering feat in itself.

That kind of stuff is readily available right now with openCV. Image detection has been very good and very fast for a long time!

Nobody really cares because most players are honest and play to have fun.

But why bother with image recognition? And pay a heavy latency penalty which makes it difficult to implement PID-style controls on player's aim? Its a game of milliseconds and opencv is comparably slow, as is the I/O involved.

Far easier to do a MiTM on the network traffic and recreate state in another process.

Still easier to clone memory pages used by the target app, or use hardware that allows dma.

In other words, tech based solutions to human problems always fails at some point probably because the problem creators can also use tech.

> We already have external “radar” cheats that use the game stereo sound to give the direction that a certain sound(such as footsteps) came from.

Did you just imply my headphones are a cheating device?

More seriously, I've heard these type of devices already exist.

I once played some Hunt (difficult, highly competitive game, a little like CS) with people on ex professional level. They obviously had gaming PCs, but they ALL had super expensive 7? Channels headphones. It's apparently a thing among ex pro (or high level) Counter-Strike players to have these super headphones that let you know where people are before regular people can hear them.

They hushed me all the time because "someone's coming". Still a fun experience - to win for a change. Don't have to add that my role was more as a decoy then anything else.

That would make an awesome competition field for AI bots that rely only on video/audio input and mouse/keyboard output.