Deepfake (hypertrucage)

A deepfake, rendered in French as hypertrucage, is visual or audio content generated or manipulated by an artificial intelligence system so as to depict a person, an object or an event in a misleading way. It engages personality rights, the criminal law governing image and voice, copyright and related rights where a work or a performance is taken over, and the transparency obligations of Regulation (EU) 2024/1689 on artificial intelligence. Article 226-8 of the French Criminal Code penalises the dissemination of visual or audio content generated by algorithmic processing which depicts a person without his or her consent, where it is not obvious that it is such content or where the content does not expressly say so.

Interference with personality rights

Depicting a person by means of a deepfake infringes the right to one’s image and, where applicable, the right to one’s voice, both of which fall within the protection of private life guaranteed by Article 9 of the Civil Code. The person’s consent, which must be specific and may be limited to a given purpose, is the sole ground of lawfulness apart from the exceptions relating to the public’s right to information. The victim may obtain, in interim proceedings, the removal of the content, a prohibition on its dissemination and an interim payment, in addition to compensation for non-pecuniary loss and, for persons whose image has commercial value, for pecuniary loss.

Copyright and related rights

Where a deepfake takes over a work, a film excerpt or a performance, it constitutes a reproduction and a communication to the public subject to the rightholder’s authorisation. The transformation of a performer’s performance, in particular the synthesis of his or her voice from recordings, engages the performer’s related right and moral right to respect for the performance. The parody exception may be invoked where the content displays a humorous intent and creates no confusion, but it covers neither attacks on dignity nor disguised commercial exploitation.

Transparency and removal

The Artificial Intelligence Act requires those deploying a system that produces a deepfake to disclose that the content has been artificially generated or manipulated, and requires providers to mark synthetic content in a machine-readable format. Removal is obtained through the notice and takedown procedure before the host or the platform, which must act promptly once informed of the unlawful nature of the content, and through the courts in the event of inaction.

Key points

A deepfake is a criminal offence where it depicts a person without consent and its artificial nature is neither obvious nor disclosed, and it gives rise to the civil liability of its author and of the person disseminating it. It may also infringe a work or a performance. Victims should promptly have the content recorded by a judicial officer, notify the platform and apply to the interim relief judge. Businesses operating generation tools must verify the consent of the persons depicted and comply with the marking obligations.

---

This glossary entry is general in scope and does not constitute legal advice.

Dhenne Avocats is a patent litigation firm. Matters outside that field, such as trade marks, designs, copyright and IT contracts, are handled by practitioners the firm works with regularly, and by the firm itself when they arise in a patent dispute.

Intellectual property attorney · Tell us about your situation