A deepfake or manipulated video can damage your reputation in a matter of minutes, as it portrays a person appearing to say, do, or approve of something that never actually happened. The immediate doubt is understandable: is it enough to report the content to the platform? Do you need to keep a copy? Does it make sense to talk about defamation, privacy, or an urgent request to a judge?
The first priority is to stop the spread without losing the evidence. I will explain which elements distinguish fake content from a simply unwelcome video, how to draft a sufficiently precise removal request, and what remedies can be used when the infringement affects your image, personal data, and reputation.
A deepfake is not just a modified video
The decisive factor is the effect the video produces on the viewer. A deepfake is an image, voice, or footage generated or altered using artificial intelligence systems to appear authentic. The European Artificial Intelligence Act specifically addresses this ability to make an unreal person, event, or statement seem real: the regulation generally imposes transparency obligations on those who disseminate artificially generated or manipulated content constituting deepfakes. However, the label alone does not make the content lawful or automatically grant a right to removal. European Artificial Intelligence Act
Technical falsity alone is not enough to define the remedy. What matters is what is attributed to the portrayed person, with what words or images, to what audience, and in what context. A clearly recognizable satirical video may raise different issues than footage published as real proof of improper conduct, a confession, or an intimate event. Furthermore, a person's notoriety does not authorize presenting them artificially in a harmful way.
The same publication can affect multiple rights. The face, voice, name, and other identifying elements may constitute personal data; the dissemination can harm reputation and image; and the accompanying text may contain defamatory statements. For this reason, it is not helpful to immediately choose a single legal label: one must precisely identify the conduct to be halted and the resulting prejudice.
Preserving evidence before removal
Removal must not also destroy the ability to prove what happened. Before submitting a report, it is advisable to note the exact URL of the content and the publishing profile, the date and time of retrieval, the title, description, relevant comments, and the visible share count. Screenshots and screen recordings are most useful when they show the page address, apparent author, and publication context together.
An isolated copy of the video may be insufficient. If the footage is downloaded without indicating where it came from, when it was viewed, and what elements accompanied it, it becomes more difficult to link it to the contested publication. In cases where the authenticity of the collection is central, technical documentation can help make files, pages, and the sequence of operations verifiable. However, this is not an automatic formula: the evidentiary value also depends on the specific dispute.
Preserving does not mean resharing. In particular, if the content is humiliating, sexually explicit, or involves minors, avoiding its further circulation protects both the person involved and the usefulness of the documentation. The material must be kept carefully and shared only with the subjects necessary for protection, without republishing it to deny or comment on it publicly.
Platform reporting and deletion requests
The report must specify the exact content and the reason for its unlawfulness. The Digital Services Act requires hosting service providers to offer accessible electronic reporting mechanisms. An adequate communication indicates the exact electronic location, i.e., URLs, explains why the content is considered unlawful, and contains the reporter's contact details and a statement of good faith regarding the accuracy of the information. Digital Services Act, Article 16
A detailed report does not guarantee immediate removal. The platform must process notifications in a timely, diligent, non-arbitrary, and objective manner; it must also communicate its decision and, where available, the possibilities for contesting it. However, there is no rule stating that any content flagged as a deepfake must be deleted within a predetermined timeframe or simply because the portrayed person denies its authenticity.
Privacy requests follow a distinct path. If the content processes personal data unlawfully, deletion can be requested from the data controller under Article 17 of the GDPR. The request must identify the material and the reason for the demand, such as the unlawful processing of data. The right to erasure has important limitations, including freedom of expression and information, legal obligations, and the defense of a legal claim: it is therefore not a power to delete any unwelcome content. GDPR, Articles 16 and 17
When urgent civil protection may be necessary
Damage that continues to spread may require a response beyond a mere private request. A formal cease-and-desist letter can demand the cessation of publication, the removal of copies controlled by the recipient, and abstention from further dissemination. This is useful because it clarifies the dispute and formulates definite requests; however, it does not compel the recipient to comply nor does it replace a judge's order.
Urgent proceedings are not automatic, but serve to prevent imminent and irreparable harm. Article 700 of the Code of Civil Procedure allows for provisional measures when the time required for ordinary proceedings would jeopardize the asserted right. In the matter of online content, the need to act urgently depends on the persistence of the video, its spread, the gravity of the accusation or exposure, and the concrete possibility that the damage will increase. Institutional reference to Article 700 c.p.c.
Removal and compensation serve different objectives. The former aims to halt or limit an ongoing infringement; compensation, on the other hand, requires proving the unlawful conduct, the damage suffered, and the causal link between that conduct and that damage. The presence of a fake video does not make the damage automatically quantifiable: relevant factors include the audience reached, online duration, defamatory messages, and demonstrable consequences on personal or professional life.
Defamation, criminal complaints, and sexually explicit content
There is also a specific offence for content falsified with AI. Article 612-quater of the Criminal Code punishes the unauthorised dissemination of images, videos or voices falsified or altered using artificial intelligence, capable of misleading others about their authenticity and causing unjust harm. The prescribed sentence is one to five years’ imprisonment. A complaint by the victim is generally required, subject to the statutory exceptions for prosecution without one. Each requirement must be established; an edited recording alone is insufficient.
Defamation may come into play if the video harms reputation in front of multiple people. Article 595 of the Criminal Code does not punish the use of artificial intelligence as such: what matters is the offense to reputation communicated to others. Footage that falsely attributes crimes, dishonorable behavior, or degrading statements may therefore require a criminal evaluation, even if the author is unknown and used a fake profile.
A criminal complaint is a formal choice rather than a simple online report. In offenses where it is required, it expresses the injured party's will to proceed criminally. The general deadline indicated by the Ministry of Justice is three months from the day the crime is discovered, barring cases where the law provides a different term. Therefore, one should not confuse the moment the video is discovered with the moment the author is successfully identified: timeliness must be considered immediately. Institutional information on reporting, formal notices, and criminal complaints
A sexual deepfake does not automatically coincide with every provision relating to the dissemination of intimate images. If real material intended to remain private is disclosed, specific rules may apply. If instead the footage is artificially created using a person's face on third-party images, shortcuts must be avoided: very serious infringements of image, dignity, reputation, and personal data remain possible, but the criminal qualification must be linked to the actual facts, not just the sexual nature of the content.
How to organize initiatives without confusing objectives
First, the content is delimited, then the tools are chosen. The platform must be able to precisely identify the video to be examined; the data controller must understand which personal data are being processed; any judicial claim requires describing the infringement and urgency; and a criminal complaint requires presenting facts deemed criminally relevant. A generic text challenging "all videos on the internet" risks being ineffective.
Initiatives can proceed together, but they do not yield the same result. Reporting to the platform seeks a decision on single content; the privacy request concerns data processing; civil protection can request injunctive measures and damages; and the criminal path aims at ascertaining a crime. Choosing one path does not always preclude the others, but each requires its own prerequisites and evidence.
The most useful fact to clarify is what the public could believe. If the video appears blatantly parodistic, one must understand whether the identification of the falsehood is truly immediate. If instead it is published as authentic proof, accompanied by accusations or reshared in thematic groups and pages, the reputational risk changes. Even a spontaneous retraction, partial removal, or the persistence of copies on different profiles affects the choice of the next step.
Frequently asked questions
Can I request removal even if I do not know who created the deepfake?
Yes, the request to the platform can concern the content even if the author is not identified. However, you must precisely indicate the URL, profile, reasons for the dispute, and any element showing why the footage is false or harmful. The anonymity of the publisher may make other remedies more complex, but it does not prevent requesting post moderation.
Does reporting to the platform replace a criminal complaint?
No, they have different functions. Reporting asks the manager to examine and limit content; a criminal complaint, in foreseen cases, expresses the will to proceed for a crime. Removing the video does not necessarily preclude evaluating the criminal profile, while filing a complaint does not guarantee on its own that the platform will immediately delete every copy.
Does the fact that the video is blatantly fake exclude damage?
No, but it affects the assessment of the infringement. Content recognizable as artificial can still be humiliating or harmful, especially if associated with a name, face, or personal information. However, if the public immediately understands its parodistic nature, the reputational scope may differ compared to footage published as authentic and accompanied by specific accusations.
Can I use the right to erasure to make every copy disappear?
The right to erasure can be useful, but it is not absolute. Article 17 of the GDPR operates when its conditions are met and encounters exceptions, including freedom of expression and information or the defense of rights. Furthermore, a request sent to one entity does not automatically eliminate content uploaded by other users or hosted on different services.
Must I wait for the platform's response before taking legal action?
There is no general obligation to wait for a response if the prejudice appears imminent. Reporting is often a useful step to quickly halt content and document the request. However, when dissemination continues and the requirements of urgency are met, precautionary protection can be evaluated without confusing the platform's waiting period with the protection of the right.
Which step to choose
The most useful choice stems from the concrete content and the speed of its dissemination. If you have identified a deepfake or manipulated video exposing you to personal or professional consequences, you can contact me to clarify which facts to document, what request to formulate, and whether prerequisites for civil, privacy, or criminal initiatives emerge.