At its SIGGRAPH 2026 keynote, Nvidia addressed backlash over DLSS 5, its upcoming neural rendering technology, after accusations that it turned games into "AI slop" by applying a heavy-handed filter. Following a controversial debut in March where demos showed altered character faces decried as "generative AI bullsh*t" by gamers and developers including New Blood Interactive, Nvidia detailed developer controls to preserve artistic intent. The core of the explanation involves two global sliders: Structure Intensity (0-1), which modulates high-frequency details like ambient occlusion, reflections, and subsurface scattering, and Tone Intensity (0-1), which modulates low-frequency details like lighting, color, and mood. DLSS 5 will also ship with three separately trained models—labeled A, B, and C—offering different interpretations of the same frame, swappable per scene. A key feature is developer masking, allowing per-object or per-character independent control of Structure and Tone intensities, enabling artists to leave a protagonist untouched by AI enhancement while applying it to the background. DLSS 5 is optimized for a single RTX 50-series GPU, down from two RTX 5090s, with VRAM requirements to be announced.