HumanProof is a verification service that certifies texts as human-authored based on clear, auditable evidence of drafting and revision.
HumanProof is currently incubated by the Media and Journalism Research Center (MJRC). However, the verification process is fully independent: assessments are carried out by paid peer reviewers, not by the submitting organisation.
All data provided for verification — such as draft files, version histories, or metadata — is handled directly by HumanProof. Private evidence is stored securely for the duration of the review and deleted within 30 days after a decision (unless the author opts in for longer retention). Only certified works are listed publicly in the HumanProof Registry.
Why HumanProof Matters
- Trust: Distinguish genuine work from AI-generated slush.
- Accountability: Give readers, funders, and editors a verifiable trail.
- Portability: A public certificate link you can embed anywhere.
Scope of Certification
At the moment, HumanProof certifies text-based works only:
- Reports
- Articles
- Policy briefs
- Research papers
- Long-form posts
Why HumanProof is Different
Most “AI-detection tools” scan text with algorithms that often mislabel content. HumanProof takes the following approach:
- We check actual documents: draft versions, tracked changes, and version history — real evidence of human authorship.
- We keep a registry: all certified works are listed in a public, searchable database.
- We guarantee independence: verifications are performed by paid human peers, not automated tools.
This makes HumanProof the first registry to combine evidence-based verification with transparent public certification.