Short answer: in most places, removing metadata from your own images is legal, and it's a normal privacy step. The legal risks come from what you remove and why: taking copyright information off someone else's work to hide infringement, removing required AI labels in China, or publishing a realistic deepfake commercially in the EU without disclosing it.

This is a plain-English overview, not legal advice. If your business depends on it, ask a lawyer in your country.

Your own images: generally fine

Metadata such as GPS location, camera serial number, dates and editing history is your information. Removing it before you share a photo is legal almost everywhere, and privacy regulators generally encourage sharing less personal data, not more.

The US rule to know is Section 1202 of the Copyright Act. It makes it unlawful to intentionally remove or alter "copyright management information", such as the author's name, the copyright notice or the terms of use, knowing that it will help cause or conceal copyright infringement. Metadata can hold exactly that kind of information.

In practice:

  • Your own work: removing your own name or notice is your call.
  • Someone else's work: stripping their author or copyright fields and then sharing the image can expose you to a § 1202 claim, on top of any infringement.

Who owns an AI image in the US

The US Copyright Office registers only work made by a human author. In Thaler v. Perlmutter, the federal appeals court in Washington, D.C. confirmed in March 2025 that a work generated autonomously by an AI system can't be registered, and the Supreme Court declined to review that in March 2026. Work where a person contributes enough creative expression, by selecting, arranging or editing, can be protected in those human parts. Metadata doesn't change who authored something.

European Union: the AI Act

The EU AI Act's transparency rules (Article 50) apply from 2 August 2026:

  • Companies that provide AI image tools must mark the output as AI-generated in a machine-readable way.
  • Businesses and professionals who publish deepfakes (realistic AI images of real-looking people, places or events) must disclose that they're AI-made.
  • People using AI for purely personal, non-professional purposes are outside the Act.

The Act puts the marking duty on the tool provider and the disclosure duty on professional deployers. It doesn't forbid an individual from removing metadata from a file, but a business publishing a realistic deepfake still has to disclose it, whatever the file says. More in our EU AI Act guide.

China: removing AI labels is prohibited

China's rules on labelling AI-generated content took effect on 1 September 2025. They require visible labels and hidden labels in the file's metadata for AI-generated content, and Article 10 forbids any organisation or individual from maliciously deleting, altering or hiding those labels. If you publish in China, don't strip AI labels from AI-generated content.

Platform rules

Even where the law allows it, platform terms may not:

  • Meta (Instagram, Facebook), YouTube and TikTok ask you to disclose realistic AI content, and can label, remove or penalise content that breaks their rules.
  • Stock sites have their own rules: Shutterstock and Getty don't accept AI-generated images from contributors, and Adobe Stock accepts them only when marked as AI.

When removing metadata is clearly fine

  • Taking GPS, serial numbers and personal details out of your own photos before sharing.
  • Removing prompts, usernames or client names from your own AI images.
  • Stopping a misleading "AI info" label on a real photo you only retouched.

When to be careful

  • The image isn't yours, and the metadata names its author or copyright owner.
  • You're publishing in China and the content is AI-generated.
  • You're a business publishing a realistic AI image of real people or events in the EU or on platforms that require disclosure.
  • You're using the cleaned image to deceive someone, such as a buyer, client or stock agency, about how it was made.

Frequently asked questions

For your own photos, yes, in most countries. Removing copyright information from someone else's work to hide infringement can break US law (17 U.S.C. § 1202).

Is it illegal to remove AI metadata from an AI-generated image?

Not in general in the US or EU, but China forbids maliciously removing AI labels, and in the EU a business publishing a realistic deepfake must still disclose it. Platform rules may also require disclosure.

In the US, not if it was generated without enough human creative input. Human selection, arrangement or editing can be protected.

No. Copyright doesn't depend on metadata. Removing someone else's copyright information can, however, create extra legal liability.