Identity & AI
Can AI use your face
without your permission?
Short answer: technically, often yes, but legally, you have rights, and increasingly there are tools to assert them. Generative models train on scraped image datasets, and the faces in those datasets rarely come with documented consent. That does not mean you have no recourse. It means the recourse has to be built.
How your face ends up in a model
Foundation models are trained on web-scale image scrapes, Common Crawl, public social media, image search. If your face has ever appeared on a public profile, news article, group photo, or tagged image, it was likely included. Once trained on, the model retains a statistical representation that can be invoked by prompts that resemble you, even without naming you.
What rights you actually have
In the EU, the GDPR treats face data as biometric data, a special category requiring explicit, informed consent. In Germany specifically, § 22 KUG protects image rights against commercial use without consent. Personality rights apply broadly: even outside the EU, most jurisdictions recognize a right against unauthorized commercial use of likeness. The rights exist; what is hard is enforcement, because you usually do not know your face was used until you see the result.
What you can do about it
Three paths exist, and they are not mutually exclusive. First, opt out where possible. Some model providers honor robots-style exclusions for training datasets. Second, monitor. Reverse-image search and identity-monitoring services can flag commercial use. Third, license proactively: by registering a documented consent record under defined terms, you convert an unmanaged exposure into a contract you control. This is what FaceLedger provides.
The asymmetry that licensing fixes
Right now, your face is being used without you knowing, without you consenting, and without you being paid. The legal situation favors you in theory and disadvantages you in practice, because finding the use, proving it, and enforcing rights is expensive. Licensing inverts the asymmetry: you set the terms in advance, you can deny entire industries (politics, pharma, gambling), you are paid per use, and disputes become contract questions rather than identification investigations.
Elk AI-gegenereerd gezicht is een juridisch risico. Tenzij het gelicentieerd is.
FAQ
Vaak gevraagd.
Can AI use my face without my permission?+
Practically, AI models are often trained on faces without per-person consent because training datasets are scraped at scale. Legally, in the EU and many other jurisdictions, this is on shaky ground, face data is treated as biometric data, and commercial use of likeness usually requires consent. Enforcement, however, is hard unless you know the use occurred.
How can I protect my identity from AI?+
Three layers: opt out where possible (some training datasets honor exclusion signals), monitor commercial use of your face, and proactively license your identity under defined terms so commercial use happens only with your documented consent and on your terms.
What if someone has already used my face in AI content?+
You may have personality-rights, image-rights, and biometric-data claims depending on the jurisdiction and the specific use. Document the use (screenshots, URLs, dates), and consult counsel. Going forward, an active license helps prevent future unauthorized use because uses on FaceLedger are filtered against your exclusions.
Why would I license my face?+
Two reasons. Control: you decide what uses are allowed and excluded, politics, religion, product categories. Compensation: you are paid per use, on documented terms. The alternative is the current default, where your face may be used without you knowing.
Klaar?
Meld je uiterlijk aan, leg je regels vast en houd de controle. Gratis voor particulieren.
Protect your identity