A suspected intimate deepfake can be harmful regardless of whether a detector confirms it. Visual anomalies alone cannot establish authenticity or generation method. Record the relevant URL safely and use the platform's appropriate synthetic-image or privacy reporting process.
Why authenticity is hard to judge
Images can be compressed, edited, retouched or synthesized in many ways. Visible artifacts can occur in genuine photos and sophisticated generated images can look convincing. Avoid treating a single visual clue or detector score as definitive. A claim about an image's origin may require specialist analysis.
Document the concern without recirculation
If the image purports to depict you, note the location, discovery date and any misleading captions. Keep a record of relevant communications. You do not need to publish comparisons with your own real images to defend yourself, and sharing the suspected image can increase its exposure.
Report harm rather than proving the model
Use the official reporting category that addresses fabricated intimate imagery or impersonation when available. Explain accurately that the content purports to depict you and why it is unauthorized or manipulated. A platform may have its own documentation requirements; use only the minimum needed.
Separate support from technical evaluation
If there is a threat, extortion or risk of serious harm, seek trusted support and appropriate authorities. A specialist may help evaluate a file if necessary, but cannot guarantee certainty. For known unauthorized material, practical removal decisions often matter more than establishing the generation software.
Step-by-step action checklist
- Preserve the known URL and relevant communications privately.
- Do not circulate the suspected image for validation.
- Choose the official synthetic-image or privacy report category.
- Retain the report outcome and seek trusted assistance.
What responsible action looks like
A fake intimate picture circulates with a misleading caption. The affected person does not post the image to ask others whether it looks real. They preserve the reported link and submit an appropriate abuse report.
Can a detector tell me with certainty a photo is AI-generated?
No. Detection may provide an uncertain signal, and real images can also contain artifacts. No single detector establishes the full facts.
