destill.ai · All sectors · Deutsch · Trust Center
For creators, creator platforms and agencies
Your video on a stranger’s channel, your voice in a fake ad. How do you show what is yours?
Creators usually face re-uploads, deepfakes and AI training alone, without a legal team. Fortress puts an invisible mark into your work at upload, so you can show later that a copy is yours and whose copy leaked. The films below are in German; the page tells the same story in English.
See also: Fortress, the engine behind it · Fortress for creators in 2 minutes · Creator business case
Deepfakes
A fake ad used an AI version of MrBeast on TikTok (source: NBC News). Your face and voice can be next.
Re-uploads
Your video turns up on another channel. YouTube publishes how many copyright removals it handles (source: Google transparency report); you still have to show it was yours first.
AI training
Whether your work ended up in a training set is hard to show. Courts are working on it (sources below); whether a mark survives into a trained model is an assumption under research, not a result.
Estimate or check: probabilistic versus deterministic
Deepfake detectors estimate
A detector looks at a finished picture and returns a probability. It can be wrong in both directions, and every new generator model makes its job harder.
Fortress SEAL checks
A photo taken live in our camera app is sealed at the moment of capture. The check compares the photo with that seal and answers yes or no: changed since capture, and if so, where. Existing or uploaded photos cannot be sealed, and the badge says "Captured live & unaltered", nothing more.
POAW: a receipt per agent step
Every step of an AI agent gets a receipt: signed with ML-DSA-65, time-stamped per RFC 3161, checkable offline with OpenSSL. It supports the evidence for the AI Act logging duties (Art. 12, 19 and 26(6)); it does not make anyone compliant.
Planned, not built: one evidence chain from the sealed photo to the decision. Pilot partners wanted.
The film
The short version, 1:30, German audio. Mia is a fictional character.
The three-minute version
2:55, German audio.
The full explainer
Chapters
What you get
Each item carries its status from our feature register: Built, Pilot or Roadmap. Hover a label for the evidence behind it.
Fortress mark at upload (RESILIENT): an invisible mark in your image that shows it is yours and helps trace a leak. It makes no statement about authenticity. Fortress TRACE: one invisibly marked copy per recipient; if a copy turns up, you see whom it was sent to. Fortress SEAL (images only): a photo sealed at live capture; after any change the seal breaks and the check shows the area. A signed manifest with C2PA export, next to the mark. - PlannedA verify API, so a platform can check a work automatically.
- PlannedA verifier as open source. The licence is not decided yet.
The legal frame and the cases
Where a mark helps you show what is yours. It does not enforce anything for you. Not legal advice.
AI Act Art. 50
Transparency duties for AI-generated and manipulated content (deepfakes), from 2 Aug 2026. The European Commission’s FAQ explains them.
Source: digital-strategy.ec.europa.eu
DSM Directive (EU) 2019/790, Art. 4 and 17
Opt-out from text and data mining, and duties of upload platforms for copyrighted works. Whether a mark counts as a valid opt-out is open.
Source: eur-lex.europa.eu
Digital Services Act (EU) 2022/2065
Notice and action on platforms. A mark helps you show which copy is yours when you report it.
Source: eur-lex.europa.eu
Getty Images v Stability AI
Judgment of the High Court of England and Wales on an image model and Getty pictures.
Source: www.judiciary.uk
LG München I (Nov 2025)
Press release of the Munich Regional Court I on song lyrics and an AI language model.
Source: www.justiz.bayern.de
OLG Hamburg 5 U 104/24
Hamburg Higher Regional Court on a training data set and the text-and-data-mining exceptions.
Source: zitierfest.de
Authors and Anthropic
Final court approval of the authors’ copyright settlement with Anthropic (report: Authors Guild).
Source: authorsguild.org
The honest limit
The mark is robust but removable, and "no mark found" does not mean "fake". The mark shows ownership and helps trace leaks; it says nothing about whether a picture is authentic. The ProofCam seal is only for live captures in our camera app, never for uploads; it says "Captured live & unaltered", never "not AI-generated". Whether traces survive into AI training is an assumption under research. A verify API, an open-source verifier and a signed manifest with C2PA export are planned, not built. Mia is fictional.
The pilot: 8 weeks, price after scoping
Scope
For a creator platform or agency: one upload flow with the Fortress mark for one group of creators, plus a test of finding the mark again after typical re-uploads. Final report with go/no-go.
Price frame: Price after scoping, based on your platform and volume.
Success criteria, agreed in writing at kick-off
- Extra effort per work step is measured; we set the target together at kick-off.
Source: our written pilot offer; the targets are what we commit to measure, not results from an earlier pilot.
What you provide
- One concrete workflow and an owner
- Test data or test images and access to a test environment
- Contacts from IT and compliance or audit
- Internet access when signing (for the time stamp)
- Your own legal assessment
- Test works from creators who agree to the test
Want the written pilot offer? Choose "Talk about a pilot" below and we send it with our reply.