Nvidia dropped DLSS 5 yesterday and called it their most significant breakthrough since real-time ray tracing. Bold claim. The problem is that when people actually saw it in action, the first thing they noticed wasn’t groundbreaking graphics. It was that every face looked like it had been processed through the same artificial intelligence filter that’s been turning Instagram into a homogenized wasteland.
Grace from Resident Evil Requiem got hit particularly hard. Her face in the DLSS 5 demo looked like someone fed her character model into Midjourney and asked for “generic video game protagonist.” The Hogwarts Legacy kids? Instagram filter casualties. Even Virgil van Dijk, an actual human being with a very recognizable face, got turned into some random dude. That’s not a breakthrough. That’s motion smoothing with an identity crisis.
The AI Aesthetic Problem
I’ve spent enough time looking at AI generated content to spot it from a mile away. There’s a signature look that’s become impossible to ignore: skin that’s too smooth, features that are too symmetrical, lighting that’s too perfect. Everything gets the same treatment. Small nose, full lips, perpetually cheerful eyes, hair that looks like plastic. On their own these aren’t necessarily problems, but when every face has all of them at once, you’ve crossed into uncanny valley territory.
DLSS 5 seems to apply this exact aesthetic to game characters. It’s not just upscaling pixels anymore. According to Nvidia, the tech “infuses pixels with photoreal lighting and materials,” which sounds great in a press release but in practice means replacing carefully crafted art direction with algorithmic homogeneity.
The thing that gets me is the industry response. Todd Howard from Bethesda said DLSS 5 lets “artistic style and detail shine through without being held back by the traditional limits of real-time rendering.” Jun Takeuchi from Capcom called it an important step in pushing visual fidelity forward. These are major voices in the industry, and they’re either not seeing what everyone else is seeing or they’ve decided it doesn’t matter.
When Hardware Dictates Art Direction
Here’s what really bothers me as someone who cares about both the technical and creative sides of game development. We’re watching a hardware company with massive market influence push a technology that fundamentally alters artistic intent. And because Nvidia dominates the GPU market, developers feel pressured to support whatever they ship.
Sure, DLSS 5 will be optional. It won’t launch until fall. It’ll require high-end hardware. Bethesda even followed up saying their art teams will be “further adjusting the lighting and final effect” for each game. But none of that changes the trajectory we’re on.
The indie developer response has been swift and brutal. Memes everywhere. Explicit statements against using generative AI in games. Some studios are even marketing themselves as “AI free” now, which is both encouraging and depressing because apparently that’s become a selling point we need.
Meanwhile the people making decisions at larger studios are lining up to praise technology that makes their carefully designed characters look like everyone else’s carefully designed characters. The disconnect is staggering.
The Slop Infiltration
This isn’t happening in a vacuum. The same aesthetic that DLSS 5 is bringing to games has already infected YouTube thumbnails, Instagram feeds, stock photo sites, and increasingly traditional media. I haven’t seen a single AI-generated film that looks good, but they keep getting made, and you can identify them instantly. That visual signature is everywhere.
Games were supposed to be different. Games have art directors and character designers and entire teams dedicated to making sure every visual element serves the creative vision. But when you layer DLSS 5 on top of that work, it doesn’t enhance the vision. It replaces it with the same smoothed-over, over-lit, generically attractive aesthetic that’s making everything else look the same.
The timing is particularly brutal given the state of the industry. Studios are closing. Thousands of talented people have been laid off. Expensive bets on live service games and endless growth have collapsed. And now here comes technology that promises to make rendering easier but at the cost of replacing human artistic choices with algorithmic ones.
I keep thinking about that comparison to motion smoothing. That’s the feature on TVs that makes movies look like soap operas, and it’s universally hated by anyone who cares about film. But it ships enabled by default because TV manufacturers think it looks “better” and most people don’t know how to turn it off. The article suggests that being a good friend might mean turning off DLSS 5 when you visit, just like motion smoothing, and that’s probably going to be accurate.
The question isn’t whether DLSS 5 will get adopted. With Nvidia’s market position and support from major publishers, it probably will. The question is whether we’re okay with letting hardware companies dictate aesthetic standards across an entire medium, and whether the homogenization of visual style is worth whatever performance gains the technology provides.