Insights
Deleting every photo of one person: a search problem first
A photograph in which someone can be recognized is personal data under the GDPR and the Swiss FADP. When that person asks a team to remove their images, the deletion takes seconds. Finding the images is the work, and it depends on whether anyone wrote down, when the shoot came in, who was in the picture.
Mathias Buschor
Co-Founder at moodcase
•
2
min read

An erasure request asks a team to delete every image of one person. It can be enforced, and what it asks of the library is a query few libraries were built to answer: search runs on what was written down, and a person appears in the pixels.
The law treats a recognizable face as personal data
The GDPR defines personal data as any information relating to an identified or identifiable person, and a photograph of someone who can be recognized falls inside it. The revised Swiss Federal Act on Data Protection (FADP), in force since 1 September 2023, applies the same test. Under Article 17 of the GDPR the right to erasure applies when consent is withdrawn or the original purpose has ended, and the controller has one month to respond.
The FADP lets the person demand deletion as well. The deadline is fixed in the text. The time the search takes is not.
An erasure request is a retrieval task
The outcome of an erasure request is fixed at intake, by whoever described the images. Only the headshot was filed as a picture of that person. Group photos, fair stands and backgrounds were filed as an event or a campaign. If the people in a shoot were named in the keywords, the request is one search. If not, someone scrolls through years of event folders, and the frame nobody recognized stays while the request is reported as done.
Automatic tagging can describe a frame as a portrait or a group. It cannot say who. That fact can only come from a person who knows, while the shoot is fresh, and the record of names has to be protected too. A library that records no one looks discreet until the first request arrives. Then it cannot comply.

What moodcase does with this
In moodcase Visual Asset Management, Asset Metadata keeps a description and keywords on each asset, and Global Search finds an asset by what is written there. AI Tagging narrows a search to portraits and groups. It does not supply the name, which stays a human judgment, kept in a workspace hosted in the EU, with a data processing agreement available.
Where this fits
The moment is a removal request that threatens to become a week of scrolling, because nobody wrote down who stood in the shoot. In moodcase the names go into the keywords when the shoot comes in, and the request becomes one search across the workspace.
Visual assets need more than a folder. See how moodcase handles the full workflow.
Try all features for 7 days. No credit card required.
Metadata
Search
Scattered Assets
Marketing Teams
Creative Teams
RELATED

Metadata is written by the person who needs it least
Tagging fails from a lack of foresight, not a lack of effort, so asking people to try harder does not help.


AI tagging: what it structures, what it does not decide.
Accuracy is a retrieval question. Authority is a different one, and no better model answers it.
