Earlier this week, Nvidia debuted DLSS 5, a new generative AI upscaling technology that purports to allow developers to get even closer to realizing their artistic visions by putting weird AI porn slop faces on popular characters like Grace from Resident Evil Requiem. People got mad. Real mad. Now Nvidia is responding by doubling down on slop-scaling.
CEO Jensen Huang was asked about the online backlash at this week’s GPU Technology Conference by Tom’s Hardware. “Well, first of all, they’re completely wrong,” he responded, arguing that full control of the tech remains with the game developers.
“DLSS 5 fuses controllability of the of geometry and textures and everything about the game with generative AI,” Huang said. “It’s not post-processing, it’s not post-processing at the frame level, it’s generative control at the geometry level.”
He claimed developers who are on board with the tech, including at Capcom and Bethesda, can “fine-tune the generative AI” to shape their game’s visuals how they wish. DLSS 5 adds a new generative AI-fueled layer of fidelity but doesn’t take away “artistic control.”
Reporter: Some are concerned NVIDIA’s DLSS 5 will make games look worse or homogeneous.
In our Q&A today, Jensen Huang responded:
“They’re completely wrong… you can fine tune the generative AI to your artistic style… if you want cartoon, toon shader, made of glass… it’s up… pic.twitter.com/cTjzUnZoiW
— Justin Ryan ᯅ (@justinryanio) March 17, 2026
Huang may be responding to the way most people have interpreted and digested the tech online, which is essentially as a sort of “slop filter” that adds an uncanny layer of hyper-fidelity to games based on generic training models rather than the vision of the original artists.
That’s due at least in part to how demo videos show before and after comparisons that make characters’ faces in games like Starfield and Hogwarts Legacy look like yassified stunt doubles. Huang’s argument is that this is being calibrated at the developer level rather than via a post-processing algorithm.
I’m not sure that logic will make much of a difference to any of the fans currently horrified by the tech’s early results. After all, what’s worse? DLSS 5 as a cheap slop-face filter or DLSS 5 as a fusion of generative AI into the fundamental geometry that defines how a game looks and feels?
DLSS 5 might do wonders for more generic details like environmental lighting, but the tools are by their very nature somewhat random and unpredictable. Most people’s only question so far about DLSS 5 is whether Nvidia can guarantee an option to keep it turned off.