Ethics policy
Last updated: 15 October 2026
Detecting AI-generated content affects real people. These are the principles we follow when building SheenProof and the limits we ask our users to respect.
1. Our purpose
SheenProof exists to help people tell what is real from what is generated, so that trust in images, video and documents can be protected. We want the tool used to support honesty, not to harm people.
2. Honest about uncertainty
Our results are likelihoods, not verdicts of fact. We show a probability and the evidence behind it so you can check the reasoning yourself. We will not claim certainty we do not have.
3. Humans stay in the loop
A SheenProof result should support a human decision, never replace it. We ask users not to take serious actions about a person, such as accusing, firing, denying a claim or reporting to authorities, based only on a score.
4. Fairness and error
Detectors can make mistakes, and some real content, such as heavily edited photos, scans, translated text or work by non-native writers, can be wrongly flagged. We work to reduce such errors, test our models on varied content and treat false positives as a serious problem, not an acceptable cost.
5. Privacy and data minimization
We use files only to analyze them, delete them after analysis and keep only the report. We aim to collect the least data needed to do the job.
6. Uses we do not support
- Targeting, harassing or surveilling individuals.
- Profiling people or making automated decisions about them without human review.
- Uploading content that is unlawful or that you have no right to analyze.
- Presenting a SheenProof report as legal or forensic proof.
- Using results to attack the credibility of real content that you know to be authentic.
We may suspend access for uses that go against these principles.
7. Transparency about how it works
Our analysis combines file metadata and content-credential checks, trained detection models and an AI reviewer (Claude) that explains the evidence. Each report lists the checks that were run and what they found.
8. Accountability and corrections
If you think a result is wrong, or you have a concern about how the service is used, tell us at hello@sheenproof.com with the report ID. We review reports of errors and misuse and use them to improve the service.
9. Keeping up with change
Generative AI changes quickly. We update our methods over time, and we will update this policy as our understanding and the law develop.