Stable Diffusion and Non-Consensual Intimate Images: What You Need to Know

Stable Diffusion is an open-source image generation model that has been used extensively to create non-consensual intimate images. Because it is open-source, it can be run locally, generating content that bypasses cloud provider content policies. This creates specific challenges for detection and discovery. However, ScanErase can help you find where your face appears in content generated by Stable Diffusion.

How Stable Diffusion is used to generate NCII

Stable Diffusion and its derivatives (including community fine-tuned models) are used to generate intimate imagery. These models can run entirely offline, making it difficult to track and discover generated content.

Why open-source models are harder to track

Commercial AI providers have content filters that block intimate image generation. Open-source models can be modified to remove these filters, and running locally means there are few logs or enforcement mechanisms available. This is why discovering content on distribution platforms is essential.

Finding Stable Diffusion-generated content

Stable Diffusion output lacks watermarks or metadata indicators. Biometric face scanning is a reliable way to find content generated from your photos using open-source models.