Why many people dislike NVIDIA DLSS 5 (but will eventually come to appreciate it)

      Upscaling, or real-time frame reconstruction for video games, is a topic of significant debate. Purists are resistant to it, but those with less powerful or mid-range gaming systems enjoy the improved smoothness it offers. NVIDIA is involved in this practice, as are AMD and Intel. However, chaos erupted when NVIDIA revealed the latest version of its super-sampling technology, primarily due to the overly AI-generated appearance of visuals, particularly human faces.

      DLSS 5 Off (left) vs DLSS 5 On (right) NVIDIA

      It has been an eventful few weeks in the tech industry, and if you've been keeping up with the DLSS 5 (Deep Learning Super Sampling) saga, you know it's been filled with reactions ranging from excitement to confusion, and even disdain. Here’s a recap of the DLSS 5 drama, moving from initial hype to the current “2D filter” state.

      The Current Situation: The “GPT Moment” That Wasn’t

      The controversy began when Jensen Huang took the stage at NVIDIA’s GTC 2026 event and announced DLSS 5. NVIDIA claimed it had advanced beyond mere pixel upscaling to generatively reimagining them. Jensen referred to it as the “GPT moment for graphics,” pledging that AI would take on the challenge of achieving visual realism, addressing aspects like skin texture, fabric shine, and complex lighting. However, the enthusiasm quickly faded.

      Awful actually😑 The more you look at it the worse it gets. This is basically Nvidia DLSS 5 applying its AI photorealism into existing games. This is taking the artistry out of the game. For example: Resident Evil Requiem from Digital Foundry pic.twitter.com/qgoK4itztF— SpawnYaard 🎮 (@SpawnYaardReply) March 16, 2026

      Within a few hours, the internet was ablaze with comparisons of Resident Evil Requiem and Starfield. The community's consensus? “AI Slop.” Rather than enhancing the games, DLSS 5 appeared to be overly softening gritty skin textures, adding unwanted makeup, and giving characters the look of 2022 Instagram influencers.

      So ... it turns out that those studios that “created” the DLSS 5 clips were just as surprised as we were. It was evidently Nvidia, which is why they all have a similar appearance. pic.twitter.com/Ubg3iECF6p— Beyond FPS (@beyond_fps) March 18, 2026

      Following that came the “Betrayal.” As reported by Insider Gaming, major game developers were caught off guard. Artists at Ubisoft and Capcom reportedly learned about the DLSS 5 demos simultaneously with the public. NVIDIA sought to manage the fallout, offering a “Full Creative Control” SDK with intensity sliders. However, the real shocker emerged just days ago: An email interview between YouTuber Daniel Owen and NVIDIA's Jacob Freeman revealed that DLSS 5 isn't actually utilizing the game’s deep 3D geometry. Instead, it’s merely a high-end 2D post-processing filter applied on top of the visuals. What was touted as a “Neural Revolution” turned out to be just a costly overcoat.

      Why “Better” Isn’t Necessarily Better

      In theory, DLSS 5 seems remarkable. And in certain aspects, it is. When viewing landscapes or static settings, AI-enhanced shadows and highlights look objectively “cleaner.” However, that’s where the issue lies: cleaner doesn’t always convey the intended atmosphere.

      Video games are a form of art, and art conveys intention.

      If a developer has invested three years crafting a dark, moody, claustrophobic hallway in a horror game, they do not want an AI interfering and “correcting” it.

      NVIDIA

      DLSS 5 tends to brighten dark areas and erase atmospheric fog because it perceives these elements as “errors” needing adjustment. The fact that developers were taken by surprise by the demo is a significant warning sign. It exemplifies typical corporate dynamics: higher-ups agree to NVIDIA for marketing appeal, while the creative teams are left uninformed. Had NVIDIA collaborated with the artists, they could have provided AI with 3D data models and blueprints.

      NVIDIA

      Consider if the AI accurately identified where a character's scar should be or how a particular fabric was meant to reflect light. Indeed, as Veedrac recently illustrated on Reddit, games with DLSS 5 that employ tone-mapping can look breathtaking. This demonstrated the technology’s potential, but it only shines when guided by human intention. By presenting it as a “black box” filter, NVIDIA effectively sidelined the very individuals who make games enjoyable.

      Then there’s the pressing issue of Data Sovereignty. As a creative designer, why would I be comfortable sharing my raw character designs and lighting maps with an AI model? We’ve witnessed the outcomes.