Five to ten times a day, I ask questions of the chatbot, Pi.AI. Whether I’m cooking, trying to remember the name of a film, or answering some common household question, it serves me well as a digital personal assistant. Though it’s currently free, I would pay to use it.

But, at its core, it’s basically just a friendlier alternative to Google. It doesn’t do something remarkably unique, or let me do something that I couldn’t do before, and it’s this way for many AI tools. They provide faster, more conversational versions of existing tools; or at least that’s true in the consumer setting. GitHub Copilot is a revolutionary tool for coding, and I know video game designers who rely on Dalle to rapidly generate concept art and add complexity to surface textures. 

But for your average user, despite the hype, AI tools don’t do that much which is completely new. There’s one exception to this, which I first experienced a few days ago when I enabled it on my Asus laptop: AI video upscaling. Available on all computers equipped with Nvidia 30 and 40 series GPUs, RTX Super Resolution works by analyzing the frames you feed it — whether watching a downloaded MP4 video or streaming on Chrome — and increases the quality by AI generating what information it feels has been lost in compression and pixelation. 

To turn it on, you just need to go into Nvidia’s outdated Control Panel software, open ‘Adjust video image settings’ and then select one of the four levels in the box ‘RTX video enhancement.’ Four is the least power intensive and One is the most powerful. 

The benefits of this are incredible.

If you’re watching some blurry old footage, this will make it look far less pixelated, with faces often looking very clear. My best example of this occurred when I tried to watch the British show, “Winter Wipeout,” for some light entertainment as we get into the Christmas season. Unable to find the program on any streaming service, or buy it from Apple, Google Play or Amazon, I instead resorted to YouTube, only to find that the only footage was in 240p resolution, iPod screen size. Stretched to full screen, even just on a laptop screen, it is an unwatchable smear of pixels. But with Super Resolution turned on, the footage is still pixelated, but the AI tool made faces clearer, along with static elements in the background and generally improved the quality. It transformed unwatchable pixelation into something eminently watchable.

The effect is even more impressive for upscaling 1080p video on a 4K display. Though it’s easy to distinguish 1080p footage from 4K, I can’t tell the difference between 1080p footage upscaled to 4K, and native 4K. If you pay for an expensive 4K Netflix plan, it’s just reduced your monthly bills.  

It’s such a simple, easy thing to do, so long as you have the right computer, and it would have been impossible a year ago, without the latest AI tools. I look forward to seeing new versions that can run on lower-end hardware.