The Future of Gaming or a “Cheap Beauty Filter”? Why DLSS 5 Is So Controversial
NVIDIA has unveiled a new version of its AI upscaler, and the first demonstration of DLSS 5’s capabilities has sparked, to put it mildly, quite heated discussions online. Unlike previous iterations of the technology, the new version fundamentally changes a game’s visuals, as it effectively redraws lighting. How the upscaler works and what exactly disappointed gamers — we break it down in this article.
What is DLSS 5 about?
The main sources of information about the updated DLSS so far are NVIDIA’s official blog post and Digital Foundry analysts, who were among the first to see the technology in action and share detailed impressions.
According to DF, NVIDIA is attempting to leapfrog several generations of GPU development using machine learning and achieve photorealistic lighting on current hardware. The new version of DLSS will run on RTX 50-series graphics cards and is expected to launch in fall 2026. NVIDIA showcased the improved upscaler using AAA blockbusters such as The Elder Scrolls 4: Oblivion Remastered and the hit Resident Evil 9. By the way, don’t miss our Resident Evil Requiem review if you haven’t read it yet.
DLSS 5 operates on the same principles as previous upscaling versions and Frame Generation, but now the technology can also reconstruct lighting, going beyond traditional rendering. In practice, DLSS 5 is effectively integrated into Frame Generation: given this new approach to lighting, it can now be said that each frame is, to some extent, “generated” rather than purely rendered using classical methods.
The neural network understands the contents of a scene — it distinguishes between skin, hair, metal, water, and processes them differently to achieve more realistic lighting. At the same time, the result remains stable even at this stage of development: the image shows no major artifacts.
The technology works with all types of graphics — from traditional rasterization to ray tracing and path tracing. The higher the source quality, the more noticeable the effect. Incidentally, we have a list of games with the most impressive RT technologies on our site: many of them may implement DLSS 5 in the future.
The most noticeable changes affect character appearance and materials: skin gains convincing light scattering, and hair looks more natural. Environments become deeper thanks to more accurate shadows and ambient occlusion. Digital Foundry experts especially praise foliage lighting, where even ray tracing previously did not always deliver satisfactory results.
It is important to note a key point — DLSS 5 works only with lighting: the upscaler does not alter textures or in-game models. However, because lighting is so critical to overall image perception, it may feel as if you are looking at two completely different characters depending on whether DLSS 5 is enabled.
At the moment, it remains unclear how significantly the neural network will impact game performance. For the demonstration, NVIDIA used two RTX 5090 GPUs at once: one ran the game itself, while the other handled DLSS 5 separately. This is a temporary solution — the technology is not yet fully optimized. The final version is expected to run on a single GPU.
It is also not yet entirely clear how much control developers will have over DLSS 5 in their games. For example, whether it will be possible to disable it for character lighting while keeping it for the rest of the scene remains unknown. NVIDIA claims that developers will be able to interact with the technology and tune it for desired results, but the extent of that control is still uncertain.
At the same time, developers will still need to work with lighting manually: even when using DLSS 5, the algorithm requires input data from the game, and as noted earlier, the higher its quality, the better the final image will be.
What do you think about an upscaler that radically changes a game’s graphics?
“They gave Grace a lip injections”
On social media and gaming forums, DLSS 5 has generally received a rather negative reaction from gamers. Users are posting parody memes mocking how characters transform when the upscaler is enabled, while also calling the feature a “cheap beauty filter” or even “an AI slop integrator for games” (mass-produced low-quality content generated by neural networks).
Most critics are primarily dissatisfied with how drastically character appearances change compared to the original. In fact, DLSS-enhanced characters often look as if they have been processed through one of many photo-editing apps that add makeup and smooth out faces. However, it is worth reiterating that nothing of the sort is actually happening here: only the lighting is being altered. At least, that is what NVIDIA claims.
Reddit is full of threads where thousands of users express their frustration. For instance, a user named TheOhrenberger, like many others, does not believe that the AI model only changes lighting and nothing else:
I was yelling at my screen when they said “it’s just changing the lighting not the assets” BRO THEY GAVE GRACE A LIP INJECTION WHAT ARE YOU TALKING ABOUT.
Another user, RareBk, drew parallels with the popular meme about a more “sexualized” Aloy from Horizon: Forbidden West:
It's genuinely the Yassified Aloy picture but real.
This tech is ridiculously terrible.
Fuck's sake the shadows are completely missing from the face in this picture.
I won't even say just the lighting bits look better, when in reality it makes the models look like bad photogrammetry.
I cannot stress that this borders on offensive for the Requiem developers because that game has so much intricate details and little developer jokes in every single texture, that an AI straight up hallucinating detail is a fucking ridiculous thing to do.
Many gamers are also unhappy not just with character visuals, but with the entire scene. For example, a Reddit user named dumahim criticized the lighting overall:
Forget the faces. Some of the shots, the shadows are just gone. Starfield, there was some dude wearing a baseball cap which cast a shadow across his face. DLSS5, the shadow is gone. Saw it with Oblivion as well. Building casts a shadow on the road below and the people nearby are casting shadows. DLSS5 on, shadows are gone (or made invisible on my phone screen I was watching on during lunch).
Even if players have no complaints about the technical execution of the upscaler, some are still dissatisfied with how radically DLSS 5 alters the visual style compared to the original, which in turn affects the game’s atmosphere. TheSphereHunter, a well-known Resident Evil YouTube creator, commented rather harshly on NVIDIA’s announcement:
Everything about this is a betrayal of these game’s artistry. painting over hand crafted intentional 3d art with shiny, wrinkly, sunken-in, porous, puckered, fraudulent filtered nonsense is deeply disrespectful. if you want this just watch gen-ai videos all day.
That said, not everyone is quick to criticize the new technology. Developers themselves appear to have reacted positively. In the same official NVIDIA blog post, quotes from several industry figures who evaluated DLSS 5 are included. For example, Todd Howard from Bethesda stated:
NVIDIA and Bethesda have a long history of pushing gaming graphics and innovation forward, and DLSS 5 represents the next major step in that journey. With DLSS 5 the artistic style and detail shine through without being held back by the traditional limits of real-time rendering. We’re excited to work with this new technology and look to bring DLSS 5 to Starfield and future Bethesda titles.
CAPCOM producer Jun Takeuchi also spoke warmly about the technology:
At CAPCOM, we strive to create experiences that feel cinematic, compelling and deeply believable — where every shadow, texture and ray of light is crafted with intention to enhance atmosphere and emotional impact. DLSS 5 represents another important step in pushing visual fidelity forward, helping players become even more immersed in the world of Resident Evil.
***
On one hand, the negative reaction from gamers is understandable: DLSS 5 does look unusual, and AI technologies in general have recently been met with resistance — especially against the backdrop of rising PC component prices. On the other hand, there is no need to jump to fatalistic conclusions. The upscaler is still being refined, which means the final version may look significantly better than the current one. Moreover, the very idea of a technology capable of altering lighting in real time is impressive, and it is encouraging that it could become a less demanding alternative to ray tracing. Ultimately, the success of DLSS 5 will depend directly on developers: if they find value in this innovation and can use it without compromising artistic intent, gamers may eventually come to appreciate its advantages.
Do you see DLSS 5 as a major threat or a potential benefit? Share your thoughts in the comments.
What do you think about character appearance with DLSS 5?
-
Developers and Players Slam DLSS 5 as a “Garbage AI Filter” That Destroys Artistic Style -
NVIDIA Announces DLSS 5: Generative AI to Reshape Gaming Visuals -
Alan Wake 2 and Control to Receive PSSR 2.0 Support on PS5 Pro -
Leaked FSR 4.1 vs. DLSS 4.5: AMD Catches Up to NVIDIA, but Stumbles on the Details -
Resident Evil Requiem Is the First Game to Showcase PSSR 2 on PS5 Pro — Screenshots Highlight the Difference -
15 Clones: Games Like Rust


