Nvidia has unveiled DLSS 5, a new PC gaming technology that uses AI to re-render video games in real time. It’s basically a make-it-realistic filter, affecting characters, foliage, textures, and lighting.
It’s but another example of how in the age of AI, the world may never be the same. And the gaming community doesn’t quite know what to think yet.
While the previous versions of DLSS simply upscaled a game’s resolution using AI, this version turns a tree that looks like a 3D model into a tree that looks like a real tree. It’s a monumental change. And a bold move. Unsurprisingly, the gaming community is fiercely divided. While some embrace the leap in visual fidelity, a loud contingent of hardcore players is furious, claiming it destroys artistic intent.
The latter camp claims it turns games into AI slop, the derogatory term that everyone loves to use now, whether it’s accurate or not. It used to mean poor-quality AI-generated images or video, but that meaning has been lost, turning into the 2026 version of “It’s Photoshop!” and “It’s CGI!” whining of yesteryear.
The way I see it, DLSS 5 looks fantastic most of the time. The intense backlash feels like it’s half posturing, and half psychological disconnect. Our brains are used to filling in the blanks of lower-fidelity graphics. And when faced with a highly detailed reality, we experience a jarring dissonance. It reminds me of how bad it feels to hear a beloved comic book character’s voice for the first time in an animated movie, and realize it doesn’t match the one in your head.
Reality bites
The two big complaints I keep hearing about DLSS are that this technology averages everyone toward one single beauty standard, and that it throws an unapproved filter over an artist’s painstaking work.
On YouTube, people like Luke Stephens called the technology “AI slop garbage,” complaining that a young character in Hogwarts Legacy “can look like a 45-year-old man.” Similarly, YouTuber AngryJoeShow lamented, “They turned DLSS into a TikTok filter. This is like hiring someone to lick off the flavor of a potato chip before you eat it.”
Critics have also lamented a loss of creative identity across DLSS-ified games. A user on Reddit argued that characters from totally different genres “all appear as if they belong to a single title, which feels like a blatant disregard for the original artistic vision.” And industry veteran John Linneman noted that while the environmental lighting shows great potential, “the character stuff is horrendous and should have been left out.”
But to me, this comment on Reddit is the one that defines the reality of it, going back to my original thesis of the brain disconnect: “The more realistic they become, the less they feel like true games.”
That’s fine. It’s a valid point. Or at least as valid as my own point of view having grown up on Atari and NES: Any game past the Game Boy Advance or the DS doesn’t feel like a true game to me. Just give me 16-bit sprites. The rest of the overproduced, stiff, motion-capture 3D-modeled stuff with infinite cutscenes you kids like is just garbage to me, not real gaming!
See? That criticism doesn’t seem to make a lot of sense, either.
The promise of AI-rendered gaming
With apologies to critics, I simply do not see how DLSS 5 is bad. Every demo I’ve watched looks perfectly in-line with the original games, showing no fundamental style change—just vastly more detail and dramatically improved lighting. Both my gamer and non-gamer friends agree that the processed graphics look superior in most instances. To me, the true uncanny valley—that creepy feeling when a digital human looks almost real but not quite—is a product of robotic motion and stiff facial expressions in big-budget games, not the sheer fidelity of the rendering itself.
Analyst Ryan Shrout rightly pointed out that “the early ‘it’s just a face filter’ isn’t the right take,” adding that the enhancements to shadows, water, and foliage are incredibly impressive. Another Redditor summarized the reality of the situation perfectly: “DLSS 5 isn’t replacing good rendering. It’s amplifying it.”
It’s also crucial to remember that Nvidia is not forcing this on unsuspecting creators. As Bryan Catanzaro, VP of Applied Deep Learning Research at Nvidia, told me back in 2023, Nvidia works closely with game developers to further develop DLSS. And remember, DLSS 5 is optional. Existing games will not suddenly morph overnight without permission; the technology requires a specific, intentional patch from the studio to even function. Gamers will also be able to opt in or out.
What we are witnessing right now is the loud birth of a new era. As Unreal Engine artist JSFILMZ observed, “It feels like we’re moving toward fully generated frames—not just upscaled ones.”
I anticipate that more developers will focus more of their energies on getting much better motion and physics and gameplay, rather than investing insane resources on extremely detailed modeling and texturing. Down the line, I can see game developers just concentrating on the game dynamics, using 3D models to guide the AI rendering precisely, which will use other types of assets (like detailed images of the characters and the rest of the game elements) to create every game frame from scratch.
So instead of panicking over a perceived loss of artistic purity, we need to recognize that our brains are simply adjusting to a digital reality that finally matches the physical one. And that we are moving into a new era, just like we went from 8- to 16- to 32-bit sprites to low-def 3D games and the current generation of PC and console gaming. The pixels are just changing again, and once the initial shock wears off, we’ll wonder how we ever played without this new tech.