Nvidia's New DLSS Technology Generates Distorted AI Faces in Game Characters

Nvidia is known for pushing the boundaries of graphic technology in gaming, and its DLSS (Deep Learning Super Sampling) system has been one of the industry's most celebrated features in recent years. However, the latest iteration of the technology is raising serious doubts among players, developers, and industry critics. The reason? The AI responsible for enhancing image resolution is generating character faces that are visibly distorted, with strange artifacts that have quickly earned the ironic nickname "AI slop faces."

What Is DLSS and Why Does It Matter?

To understand the problem, we must first comprehend what DLSS is. Nvidia's technology employs neural networks and machine learning to upscale images in real-time during gameplay. Instead of rendering the game at native resolution — which requires a lot of processing power — DLSS renders the game at a lower resolution and uses AI to "guess" and reconstruct the missing pixels, delivering an image that ideally looks native or even superior.

In previous versions, especially DLSS 2 and DLSS 3, the technology was widely praised for:

• Significantly increasing performance without noticeable loss of visual quality
• Reducing VRAM usage in high-resolution scenarios
• Making ray tracing more viable on mid-range hardware
• Providing a smoother experience on high refresh rate monitors

However, the new version introduced a significant change in how the AI interprets and reconstructs game elements — and character faces seem to be paying the highest price for this evolution.

The Problem of "AI Slop Faces"

Players who activated the new version of DLSS in popular titles began to notice something disturbing: character faces, especially in cutscenes and close moments, were displaying bizarre visual deformities. Slightly misplaced eyes, skin textures that appear "melted" at certain angles, facial expressions that look strange during quick movements, and hairline edges with flickering artifacts.

The website Kotaku, a reference in gaming journalism, was one of the first to document and name the issue, using the term “AI slop faces” — a nod to the growing criticism of visual content generated by AI that lacks quality and visual coherence. The term "slop" (mess, rubbish, something poorly made) has been widely used online to describe AI-generated images that appear incorrect, unstable, or visually disturbing.

Which Games Have Been Affected?

While Nvidia has not released an official list of problematic titles, player reports indicate that the issue is most visible in:

• Games with highly detailed characters, where any deviation in the face is immediately noticeable
• Titles with long cinematic scenes, where the upscaling system needs to maintain consistency for longer periods
• Games with a lot of action and fast camera movements, which confuse the frame reconstruction algorithm

Why Is AI Getting Faces Wrong?

The technical answer lies in how deep learning models work. DLSS utilizes a neural network trained on millions of game frames to learn how to "fill in" the missing pixels. The problem is that human faces are extremely complex and sensitive. Any small reconstruction error becomes immediately visible to the human eye — a phenomenon related to the concept of the "uncanny valley."

When the AI tries to reconstruct a moving face, it needs to:

1. Accurately identify the edges and contours of the face
2. Maintain temporal consistency between consecutive frames
3. Preserve subtle details such as expressions, pores, and reflections in the eyes
4. Handle partial occlusions (hair in front of the face, shadows, etc.)

If any of these steps fails — even minimally — the result can be unsettling. And the more aggressively Nvidia attempts to use AI to maximize performance gains, the greater the chance for these errors to appear.

Nvidia's Response and the Community

As of the publication of this article, Nvidia has not issued an official statement acknowledging the problem comprehensively. The company typically rolls out driver updates and adjustments to the DLSS AI model to fix specific visual artifacts, but this process may take weeks or even months.

In the gaming community, the reaction has been mixed:

• More technical players recommend disabling DLSS or using more conservative settings (like the "Quality" mode instead of "Performance") to minimize artifacts
• More severe critics question whether Nvidia is prioritizing performance benchmarks over actual visual quality
• Technology defenders argue that issues like this are normal in initial releases and that fixes will come quickly

DLSS vs. Competitors: The Context of the Dispute

It's worth noting that DLSS is not alone in this market. AMD offers FSR (FidelityFX Super Resolution) and Intel has XeSS (Xe Super Sampling) as alternatives. Curiously, due to their different approaches — FSR, for example, does not rely on dedicated hardware and uses more traditional algorithms — they may, in some scenarios, exhibit fewer artifacts in faces, although they generally deliver inferior performance compared to DLSS under ideal conditions.

This controversy reignites an important debate: how far can we trust AI to make critical visual decisions in real-time?

What to Expect Going Forward

Nvidia has a history of responding quickly to significant complaints from the community. It is expected that:

• Driver updates will be released in the coming weeks to mitigate facial artifacts
• Game developers may need to adjust the implementation of DLSS in their specific titles
• The company should retrain parts of the AI model focusing on facial stability

For now, the recommendation for players noticing the problem is to test different DLSS settings — particularly reducing the upscaling level — or temporarily disabling the feature in games where character visual fidelity is a priority.

Conclusion

The controversy of DLSS's "AI slop faces" serves as a reminder that artificial intelligence, no matter how powerful, still makes visible and sometimes disturbing errors. In a sector where visual immersion is crucial, any distortion in a character's face can completely break the player's experience.

Nvidia will need to act quickly to preserve the reputation of a technology that, until now, was considered one of the greatest assets of GeForce GPUs. The future of DLSS depends not only on impressive frames per second but on a visual quality that players can truly trust — pixel by pixel, face by face.