Skip to content

Nvidia DLSS 5 Is Turning Game Characters Into AI Slop

Young woman with blonde hair and leather jacket stands in a wet urban alley with two people walking under an umbrella.

Nvidia, the world’s biggest company and the business once known as ‘that graphics card lot’, has unveiled a technology it describes as an "AI-Powered breakthrough in visual fidelity for games". We would call it “a nightmarish AI effect that makes game characters look as though they were created from a Google Gemini prompt”.

Nvidia DLSS 5 and game character detail

DLSS 5 is an impressive technical achievement in several respects, as Nvidia’s newswire explains. Using compatible Nvidia graphics cards, it can “infuse” pixels on screen with photorealistic lighting and materials. Nvidia’s sample clips show the technology in action, apparently adding substantial new detail to game characters.

That is also precisely the issue: the technology seems to be literally painting over the work of the people who originally created those characters. The resulting look is utterly uniform. Grace Ashcroft from Resident Evil Requiem resembles an AI girlfriend with an ozempic prescription. When DLSS 5 is used with Starfield, Marika Boros acquires a notably similar complexion.

AI concerns and the DLSS 5 reaction

The broader context is relevant as well. By now, everyone is familiar with the AI bubble. If not, the opening line of this story - and the reminder that a PC graphics card company has somehow become larger than Apple and Meta - should serve as a useful prompt. Firms such as Nvidia have invested heavily in AI technology and are now seeking every possible way to insert it into everyday life.

High-profile developers have spoken positively about the technology. In Nvidia’s announcement, Bethesda’s Todd Howard said: “We’re excited to work with this new technology and look to bring DLSS 5 to Starfield and future Bethesda titles."

Capcom's Jun Takeuchi was equally upbeat: "It represents another important step in pushing visual fidelity forward, helping players become even more immersed in the world of Resident Evil."

That view is sharply at odds with the wider public response. Critics have compared the effect to transforming games into AI slop, while questioning what it means for the integrity of a game’s original art direction.

Jensen Huang defends Nvidia DLSS 5

Nvidia’s Jensen Huang has, however, pushed back against those objections. “First of all, they’re completely wrong,” he told Tom’s Hardware when asked about the negative response to DLSS 5.

“The reason for that is because, as I have explained very carefully, DLSS 5 fuses controllability of the geometry and textures and everything about the game with generative AI.”

Huang also said that “it doesn’t change artistic control” and that “it’s not post-processing at the frame level, it’s generative control at the geometry level".

Yet Huang’s answer appears to address the technical process behind the effect rather than the visual outcome, which is what critics have chiefly objected to. If creators continue to have artistic control, why do the demonstration clips appear so homogenous?

Expect plenty of highly cautious technical explanations - and, hopefully, games that do not resemble a Nano Banana prompt - in the near future, as Nvidia tries to take back control of the DLSS 5 narrative.

Comments

No comments yet. Be the first to comment!

Leave a Comment