New assessment
Check one video across technical, aesthetic and narrative markers, then record your decision. The file is analysed in this browser and is not uploaded.
Drop a video file to start
MP4, MOV and AVI files are checked for content credentials. WebM and MKV get metadata and measurement checks only.
What happens to the file
- Fingerprint and content credentialsA SHA-256 fingerprint ties the record to this exact file. Any C2PA credential is read and its signer checked against the C2PA Trust List.
- Metadata and measurementsCapture device, encoding software, dates and encoding, plus measured shot changes and sound levels.
- Markers that applyYou say what the video claims to be; the aesthetic and narrative markers that apply are selected, each with its reason and the file's evidence.
- Your decision and the recordYou record the decision and your reasoning, and export it as JSON or PDF.
Nothing leaves this browser unless you export the record, search a claim on Google Fact Check Explorer, or consent to the optional AI-generation check for a file.
Authenticity Lens was selected for the Global Trust Challenge prototyping phase in June 2026 by the IEEE Standards Association, AI Commons and OECD.AI, in the AI Literacy and Civic Education cluster. It is a judgement-support tool, not authentication infrastructure: it builds people's capacity to assess synthetic video and does not certify content at source.
Three marker layers are checked in the order people weigh credibility: technical, then aesthetic, then narrative. Automated findings are one layer of evidence. The person makes the call.
Working in this prototype
- Your own video, analysed entirely in the browser
- C2PA content-credential reading and validation
- Container metadata and encoding analysis
- A SHA-256 fingerprint of the exact file assessed
- Shot-change and sound-level measurement
- Relevant aesthetic and narrative markers selected for each video, with reasons for those that don't apply
- Timestamped flags and measured shot changes on the clip's timeline
- A provenance inventory, a decision record and JSON, PDF or signed export
- An optional AI-generation check, sent only with consent for each file, when connected on the server
Not in this prototype
- Analysis from a pasted link (download the file first)
- Trained forensic detection models (MSCA phase)
- Invisible-watermark detection (MSCA phase, access-dependent)
- Cross-cultural aesthetic and narrative models (MSCA phase)
- Institutional storage of assessments (MSCA phase)
This prototype is the first stage of a GTC, MSCA and commercialisation pathway. The Marie Skłodowska-Curie Postdoctoral Fellowship (proposal 101356885) with the CYENS Centre of Excellence and the Cyprus University of Technology would deliver the next stage.
Roadmap
- 2026GTC prototype: three-marker framework, human-guided interface, in-browser C2PA and metadata analysis, exportable record.
- 2027 to 2029MSCA fellowship at CYENS: codesign with educators and journalists, trained forensic and narrative models, cross-cultural evaluation in Cyprus, the UAE and South Asia, an open-access platform.
- After 2029Smart Story Labs: an enterprise platform within InsightOS, institutional licensing and API access.
Research questions
- What equips people to make warranted authenticity judgements about synthetic video?
- How can the three-marker framework become a human-centred, cross-culturally validated platform?
- How do verification skills and credibility judgements vary across Cyprus, the UAE and South Asia?
People
- Prof. Dr Andri Ioannou, primary supervisor, Cyprus University of Technology and CYENS
- Dr Aekaterini Mavri, secondary supervisor, Cyprus University of Technology
- Dr Amritha Subhayan Krishnan, applicant, founder of Smart Story Labs
- Fingerprinting the file (SHA-256)Waiting
- Reading content credentials (C2PA)Waiting
- Reading container metadataWaiting
- Measuring shot changesWaiting
- Measuring sound levelsWaiting