A sample video shows how computer vision (running on an external computer) detects an enemy and calculates how far the mouse needs to move to target that enemy.
Just a few frames later, thanks to inputs sent through external hardware, the cheat user automatically targets the enemy and fires.
Another demonstration of computer vision identifying multiple valid targets inside the central “kill zone.”
The detection seems to work quickly in a variety of lighting and obscuring conditions.
Input devices like the Titan Two are key to making computer vision cheat tools like this work.
A mock-up of the user interface for one computer vision cheat tool, showing how it can be configured for head, body, or leg shots.
Users may need to play with detection thresholds and speed settings to get the cheat tool working correctly, the maker says.
The size of the central “kill zone” can be changed, affecting the detection speed and how far the auto-aim “snap” can go.
The makers of a machine-learning-powered auto-aim cheat tool have halted development and taken down their website “at the request of Activision Publishing.” That’s according to a message posted on the User Vision website, in which developer USER101 promises the group “will no longer be developing or providing access to software that could be used to exploit [Activision]’s games. My intent was never to do anything illegal.”
Previous versions of the User Vision software offered more basic cheating functions like “recoil adjustment” that automatically steadies a player’s aim and a “trigger bot” that shoots automatically when an enemy is detected in the player’s crosshairs. Last week, though, the makers of User Vision started promoting a future version that would provide “full auto-aim [and] full auto-shots” on “any game” for PC, Xbox, or PlayStation.
User Vision’s proposed auto-aim cheat was designed to work without any modifications to the hardware or software running the game itself, thanks to a combination of external capture cards, machine-learning algorithms to detect enemies, and external hardware to emulate user input.
Many of User Vision’s promotional videos—which showed the tool being used in Call of Duty games—were taken offline late last week due to a copyright claim from Call of Duty publisher Activision. The entire YouTube account seems to no longer exist, and the User Vision Discord appears to be offline as well.
While USER101 wrote on the User Vision website that “this statement was not required,” it seems clear that legal threats from the publisher are behind the sudden shutdown (Activision has not responded to a request for comment from Ars Technica). Those threats are no joke; publishers including Riot Games, Ubisoft, Take Two, and Epic Games have taken cheatmakers to court in recent years. The cases have often resulted in significant monetary damages against software-makers that violate an online game’s terms of service.
USER101 stressed that the auto-aim functionality showcased in User Vision’s videos had yet to make it to a published version of the software. USER101 also tried to refocus the conversation by saying that “this type of technology [has] other actual assistive benefits; for example, by pointing a webcam at yourself, you could control movement without the use of limbs. Unfortunately, because of its potential negative impact, I will not be developing it further.”
While User Vision promoted its upcoming auto-aim process as “undetectable [and] unstoppable,” game developers said they already have ways of dealing with these kinds of cheating tools, in part by looking for superhuman play patterns. “Ultimately, the ’emulated input’ vector isn’t anything new, and the Vanguard team is very aware of it,” Valorant anti-cheat lead Phillip Koskinas told Ars. “Cheaters are always looking for new corners to hide in, and ‘Kernel Drivers’ have never been the most important tool in our arsenal.”
The apparent death of User Vision won’t stop the same general cheating method from continuing in other corners, of course. As artificial intelligence technology gets more powerful and accessible, anti-cheat efforts will need to be increasingly vigilant to detect non-human play of this type.