Deepfake discovery is a technology arms race. Discovery tools achieve high accuracy against known AI generation models but consistently lag behind the most recent outputs. You do not need to prove that content is AI-generated to assert your rights. ScanErase focuses on finding exposures of your face across the internet.

Available deepfake discovery tools

Professional tools include Microsoft's Video Authenticator, Intel's FakeCatcher, and Sensity AI's commercial platform. Research tools from universities provide higher accuracy at the cost of accessibility. Consumer-facing tools like WeVerify and Deepware Scanner offer web-based detection with lower accuracy.

Limitations of current technology

Current discovery tools struggle most against diffusion model-based generation (the type used by most consumer nudify apps). Compression artifacts from downloading and re-uploading also confuse discovery algorithms. No tool is 100% reliable.

When discovery results are uncertain

If a discovery tool returns uncertain or negative results on content you believe is AI-generated, this does not affect your rights. Assert your rights based on the non-consensual nature of the content, not on technical determination of AI generation.

Frequently asked questions

Which deepfake discovery tool is most accurate?

Intel's FakeCatcher and professional-grade tools from Sensity AI achieve the highest accuracy rates against known models. Accuracy drops significantly against the most recent diffusion model outputs.

What if a discovery tool says content is not a deepfake but I believe it is?

Discovery tools are not infallible. You can assert your rights based on the non-consensual nature of the content regardless of discovery tool output.

Related deepfake resources