What Is Holding Back Eye Tracking in Video Games?

Eye tracking is one of those sleek new technologies that seemed exciting at first, but never really took off in the wider consumer market. It promises a more natural and intuitive means of interaction with modern technology, and what better place to introduce intuitive UI than video games? One company in particular that has done its best to bring eye tracking to gaming is "tobii". They sell two options for hardware that enable you to use a variety of eye tracking software, including a surprising number of integrations. As of April 2020, there appear to be: 152 eye tracking enabled games, Windows 10 integration, and Mobalytics, a subscription service for improving your League of Legends gameplay. But how is eye tracking actually used in games?

You can find clips of these "immersions" being used in a handful of eye tracking enabled games (from the list linked above). Without trying it for myself/yourself, it seems hard to say which of these actually make gameplay more immersive. But there are a variety of enhancements available, and an assortment of games with integration, so I can only assume that some of these actually succeed at improving the gameplay experience. Is it such a small improvement, though, that eye tracking has remained this unpopular? Or is there something else holding it back? I believe that there are 3 main barriers to wider adoption of eye tracking in video games. They do not all need to be overcome, nor are they equal in importance, but each one holds this technology back in one way or another.

1. Hardware Cost

tobii's two hardware options are each over $100. This is a steep ask for customers (other than those with accessibility needs) when the features that come along are generally quality-of-life, or otherwise seem to be small improvements. But almost every modern laptop has a built-in webcam, and webcams are very common among desktop computer users. So if we all have a camera ready to go, many customers may be wondering why they need to buy more hardware at all. And as a matter of fact, they don't! Free browser libraries such as WebGazer.js exist which can use your simple webcam. But there is a tradeoff...

2. Calibration

Eye tracking always requires some form of calibration. This typically involves some user input, such as self-calibration by matching mouse clicks to gaze location on the screen. But going through these steps would only be a small nuisance, and some eye tracking can handle calibration automatically. The real issue faced here is the quality of the calibration. Especially with libraries that use built-in webcams, accuracy is flaky and inconsistent, making many use cases quite inconvenient.

3. Being Watched...

This is very different from the first two barriers, because it is more of a social barrier than a technological one. For certain users, there will always be a lurking concern that they are being watched through any live camera connected to their computer. Webcam covers are a popular peripheral for precisely this reason.

If the reliable hardware options become cheaper, or reliability improves for software-only options, then we'd be one large step closer to wider adoption of eye tracking in video games. There is also potential for a social shift of perspectives on surveillance to affect the popularity of this technology (though this could be positive or negative). Either way, I'm interested to see whether eye tracking in video games has already peaked in popularity, or if it has more potential in the years ahead.