Nvidia DLSS 5. A Milestone or a Massive AI Failure in Gaming? – AI News – #4 March 2026

For years, Nvidia has been pushing the boundaries of what can be generated on our monitors. From the debut of real-time ray tracing in 2018 to increasingly advanced upscaling methods. According to Jensen Huang himself, the company’s CEO, the latest installment of their flagship product is meant to be the biggest breakthrough in computer graphics in recent years.

On the other side of the barricade, however, stand gamers and reviewers who, upon seeing the first shared video materials, noticed something entirely different—a loss of depth, detail, and artistic soul in favor of a smoothed-out, plastic image straight out of cheap mobile app filters.

In the ideal world painted by the “green team’s” PR departments, DLSS 5 is a bridge between the limitations of real-time rendering and Hollywood special effects, which take days to generate.

Instead of merely boosting resolution or generating additional animation frames (as previous versions did), the algorithm analyzes input motion vectors and colors, and then uses artificial intelligence to infuse the scene with fully photorealistic lighting and material properties. According to Nvidia, the model can understand scene semantics—it distinguishes human skin, material textures, or the way light penetrates hair. It sounds like an absolute triumph of machine learning in the service of gaming. Support for the technology has already been declared by industry giants such as Bethesda, Ubisoft, and Capcom.

It quickly turned out, however, that what sounds beautiful in press releases does not necessarily work well in the dynamic environment of video games. The community started asking questions, and experts began their technical analysis.

Initially, it was suggested that DLSS 5 deeply integrates with the game engine, understanding the 3D geometry of objects. However, tech journalist Daniel Owen decided to dig deeper and obtained answers from Nvidia that shed a completely different light on the matter. It turns out that DLSS 5 operates mainly based on flat 2D screenshots enriched with motion vectors.

In practice, this means we are dealing with a locally running AI filter that instantly “paints over” consecutive frames. Such a mechanism is a simple recipe for so-called AI hallucinations and a mass of visual distortions, especially in productions characterized by fast, dynamic action. What is worse, developers lose a significant part of their control over the final image—they can only adjust the intensity of the effect or mask selected objects.

Technical issues are just the tip of the iceberg. The biggest problem with DLSS 5 seems to be how it affects the artistic direction of games. A great example is the latest installment of the cult horror series—Resident Evil Requiem. In the showcased footage, we saw how DLSS 5 treated the character of Grace Ashcroft. Instead of a face full of details and wrinkles, we received an unnaturally smoothed, “plastic” version of the protagonist. The game gained an artificial gloss, losing its dark atmosphere meticulously designed by the artists.

Artistry in games often relies on flaws—on dirt, mimic wrinkles, or specific, not entirely realistic, but atmosphere-building lighting. Handing the brush over to an algorithm that averages the image based on its training data met with massive resistance from players, who bluntly called this effect “AI-generated slop.” The pressure proved so great that studios like Bethesda have already had to salvage the situation, issuing statements that in upcoming updates to Starfield, this technology will be completely optional and will remain under the constant control of human artists.

The introduction of DLSS 5 is undoubtedly a technical milestone in the development of artificial intelligence and its real-time implementation. It is hard to ignore the computing power behind this endeavor. However, at this moment, from an artistic and image perspective, it is a massive failure. The community’s concerns are fully justified—there is a real risk that game publishers, looking for savings, will begin to marginalize the work of true art directors and graphic designers, leaving the visual side of games in the hands of algorithms lacking any artistic sensibility. Time will tell if Nvidia takes this lesson to heart and gives creators tools that allow for actual control, or if we are facing an era of “perfectly smoothed” but soulless virtual worlds.

Want to stay up to date with the most important changes in the world of AI and SEO? Technological trends change from week to week, and what was a revolution yesterday might turn out to be a dead end today. Don’t let vital information bypass your industry and slip right past your nose! Sign up for the Delante newsletter—we will deliver reliable analyses, the latest news, and proven advice straight to your inbox. Join us right now!

Sources: https://nvidianews.nvidia.com/news/nvidia-dlss-5-delivers-ai-powered-breakthrough-in-visual-fidelity-for-games + https://www.ign.com/articles/nvidias-dlss-5-is-a-slap-in-the-face-to-the-art-of-video-game-design