So, Nvidia announces DLSS 5a new AI graphics technique that uses generation systems to “enhance” video games with more photo-realistic effects. I’m not going to worry about mincing words here. I think it looks crap. Yes, we’ve never seen even a minute of it in action, but if what’s being teased is the direction the tech giants are thinking about the future of graphics technology in gaming, I’m afraid I’m out.
The first shot of Nvidia’s DLSS 5 announcement trailer gives us a good idea of the technology’s impact on Capcom’s latest title, Resident Evil Requiem. It’s already a great looking video game, so I can’t say I’ve ever felt the need for improvement, but what’s surprising is that every time a green bar crosses the screen, we find a grumpy Grace Ashcroft staring back at us without any discernible features, as if the light behind her eyes has been snuffed out by technology. It’s the kind of smooth face and unrealistic lighting we’re used to seeing in the corners of app stores and advertising banners on websites we only see in incognito mode. It takes a character that was very carefully crafted by Capcom’s art team and adds a layer of gloss that says, “No, we can do better than that,” and makes Grace stand out within Requiem’s world, rather than feeling like she’s part of it.
I’ve never played Resident Evil Requiem and felt like it wasn’t realistic enough or that either of the two main characters needed glow-ups, and I’ve never played an EA FC 26 match where I wished Virgil Van Dyke didn’t resemble his real-life opponent. I play games to experience finely crafted artistry, whether the developer is trying to transport me to a faraway fantasy world or recreate the real world as faithfully as possible. But DLSS 5 doesn’t offer me any of that, instead the AI drips a large container of oil onto the canvas instead of a human-held paintbrush. what are we doing here?
AI has no artistic or authorial intent. It reads the image as if it were purely 0’s and 1’s and overwrites it according to the training data, theoretically to make the image look “better”. In an accompanying Nvidia blog post, the company explains that the model is trained to “understand the semantics of complex scenes such as characters, hair, fabric, and translucent skin, along with ambient lighting conditions such as frontlight, backlighting, and cloudy conditions, end-to-end, all by analyzing a single frame.” In theory, the idea is to improve the appearance of the characters while keeping them rooted in the scene they’re already standing in. However, the result was unpleasant to my eyes. Each Hogwarts Legacy character in the trailer appears to be spotlighted from behind the camera, but it never feels natural. Yes, we now live in a world where the majority of game environments are dynamically lit, but the developers and technical artists behind those systems still have ultimate control over how they look. They get to decide the mood they’re trying to set, and they spend a lot of time making sure it fits their vision for the game, but Nvidia and the engineers behind this AI filter clearly think they know better.
Art direction is a huge part of video game design. The worlds and characters these developers have hand-crafted over the years are what ground our experience. I just recently started replaying Uncharted 4, and I’m still amazed at how incredibly subtle Nathan Drake’s face is during cutscenes in this nearly 10-year-old game. Small wrinkles, scars, and bruises appear and disappear throughout the story, reflecting his place in the world and the struggles he is experiencing. I couldn’t imagine wanting to layer an AI filter on top of it. It will no doubt smooth out his wrinkles, remove the blemishes, and recondition Naughty Dog’s flawed hero to better reflect the “perfect” man promoted by society and flooded with training data. But cuts, scrapes, and genetic “flaws” are the small details that connect us to the characters and are the intentions of the artists who created them.
The technology behind DLSS 5 not only makes the game visibly distracting, but if adopted by the companies that employ these artists, it could completely change the feel of the story. I can only imagine the vast majority of video game developers let out a collective sigh when this trailer was released, but I’m afraid those in charge of the funds might have smiled a little instead.
This feels like the beginning of a new era, and it’s a story that continues far beyond Nvidia’s announcement this week. Already, we’re seeing backlash from fans slamming DLSS 5 as an “AI-generated blunder.” Bethesda promised to make “further adjustments to the lighting and final effects” soon. Starfield introduced the technology after its space RPG showed multiple characters suffering the same uneventful fate. “This is all under the artist’s control and completely optional for the player,” Bethesda Game Studios added on social media. While it may be completely optional for existing games for now, I’m concerned about what will happen when studios are forced to use this technology to fit their development processes.
If we allow this kind of technology to flourish, will we give companies the green light to de-emphasize the importance of well-chosen art direction and instead do the bare minimum and let AI fill in the gaps?I don’t know about you, but I like my art to be created by humans. I’d like to know if someone decided to light the scene a certain way, or if the details on the characters’ faces were sculpted with purpose. So I keep saying that visual “upgrades” like this look like shit. In any case, the technology behind it does not have any harmful sentiments.
Simon Cardy is a senior editor at IGN, where he mostly spends his time wandering around open-world games, indulging in Korean movies, and despairing over the state of Tottenham Hotspur and the New York Jets. Follow him on Bluesky. @cardy.bsky.social.
