Sexual Deepfakes: Criminal Defence Against AI-Generated Pornographic Montages

Sexual deepfakes —pornographic montages generated with artificial intelligence that insert a real person's face or body into sexual content without their consent— have become one of the most complex fronts of technology criminal law. The difficulty is twofold: on the one hand, the reputational and moral harm to the victim is immediate and viral; on the other, the Spanish Criminal Code (CP) does not yet contain a specific offence punishing "the sexual deepfake" as such. As criminal defence lawyers specializing in sexual deepfakes, we explain through which offences this conduct is prosecuted, the penalties, the doubts over classification and the lines of defence.

There Is No "Sexual Deepfake Offence"

It is worth starting from an essential premise: Spanish criminal law does not autonomously typify the sexual deepfake. There is no article stating "whoever creates or distributes a pornographic montage generated with AI shall be punished with...". What exists is a set of pre-existing offences that, depending on the content of the montage, the means of distribution and the age of the person depicted, may apply individually or concurrently.

This has first-order practical consequences. The legal classification of the facts is not automatic but depends on a careful analysis of each case. And where there is room for classification, there is room for defence: arguing which offence really applies —or whether any does— is one of the lawyer's central tasks.

Moral Integrity (Art. 173 CP)

One of the most frequently used routes is the offence against moral integrity under Art. 173 CP, which punishes anyone who inflicts degrading treatment on another person, seriously undermining their moral integrity, with six months to two years in prison. The creation and distribution of a pornographic montage that exposes the victim as a sexual object without their consent may amount to that degrading treatment, especially where the content goes viral and seeks to humiliate them.

Moral integrity protects the dignity of the person against objectification and debasement. In sexual deepfakes, the reproach does not fall on a physical assault but on the instrumentalization of the victim's image in order to subject them to public sexual ridicule.

Privacy and Intimate Images (Art. 197.7 CP)

Art. 197.7 CP punishes the non-consensual disclosure of intimate images or recordings obtained with the victim's consent in a personal or domestic setting, where their disclosure seriously undermines their privacy. The penalty is three months to one year in prison or a fine of six to twelve months; whoever, having received them, re-distributes them faces a fine of one to three months. The penalty is imposed in its upper half where the victim is a minor or a partner, or where there is a profit motive.

Here arises one of the most debated legal questions. The provision is designed for real intimate images previously obtained with consent (the classic scenario of a partner who circulates private photographs after a break-up). In a purely synthetic deepfake there is no pre-existing real intimate image, but an artificial creation. For that reason, the application of Art. 197.7 CP to content generated entirely by AI is disputed: part of the legal scholarship considers that the offence requires a real image obtained with consent, a requirement the synthetic montage does not meet. This debate is not theoretical: it may be decisive for the defence.

Honour and Defamation (Art. 208 CP)

Art. 208 CP defines injuria (defamation) as an action or expression that harms another person's dignity, undermining their reputation or attacking their self-esteem; only serious defamation is criminal. Where the sexual deepfake is distributed with publicity, serious defamation is punished with a fine of six to fourteen months (Art. 209 in relation to Art. 208 and Art. 211 CP, which governs publicity).

This route becomes especially relevant where the victim is a public figure or where the montage seeks to attribute specific sexual conduct to them in order to discredit them. Unlike offences against privacy or moral integrity, defamation is usually prosecutable only at the victim's instance, which conditions the procedural strategy.

Virtual Child Pornography (Art. 189 CP)

The most serious scenario arises where the montage depicts a minor or a person with a disability in need of special protection. Art. 189.1 CP punishes child pornography with one to five years in prison and expressly includes virtual or apparent pornographic material: realistic images of a minor generated by any means, even where the minor does not actually exist or did not take part.

This means that a deepfake realistically depicting a minor in a sexual context may amount to child pornography even if it is entirely synthetic. It is the harshest criminal framework of all those applicable to sexual deepfakes, which is why the apparent age of the person depicted becomes a decisive element, for both prosecution and defence.

The European Framework and the Pending Reform

Beyond the Criminal Code, it is worth setting the regulatory context. The Digital Services Act (DSA, EU 2022/2065) imposes content-removal obligations on platforms regarding illegal content, and the AI Act (EU 2024/1689) establishes, in Art. 50, transparency and labelling obligations for artificially generated or manipulated content. These are, however, obligations of an administrative nature, not criminal offences.

On the criminal side, there is an artificial intelligence bill introduced in May 2026 that envisages amending the CP to address non-consensual sexual deepfakes specifically. Until it is passed and comes into force, it is not applicable law and can only be mentioned in conditional terms. Any advice must be based on the legislation in force, not on announced reforms.

Computer Forensic Expert Evidence

In these matters, digital evidence is the terrain on which much of the proceedings is decided. Defence and prosecution work with elements such as:

• Metadata of the file and its creation, which may reveal the date, the device or the tool used.
• Identification of the AI tool used to generate the montage and of the traces it leaves in the content.
• IP traceability and the accounts from which the material was created or distributed, with their limitations: an account or an IP address do not, by themselves, identify a natural person.
• Chain of custody and distribution: how the material was obtained, copied and brought into the proceedings, and through which channels it spread.

Challenging the lawfulness and integrity of this evidence, as well as its chain of custody, is one of the most effective tools of the defence.

Lines of Defence

Faced with a charge of creating or distributing a sexual deepfake, the main lines of defence are:

• Lack of material authorship. Contesting that the accused is really the person who created or distributed the content, where only technical indicators exist (an IP, an account) that do not conclusively identify them.
• Satire or parody. In certain contexts, especially with public figures, the montage may seek to rely on freedom of expression as satire or parody, which requires an analysis of the content and its purpose.
• Mistake as to the nature of the material. In cases of re-distribution, it may be argued whether the person knew of the sexual nature or the lack of consent attached to the content.
• Argument over classification. Since there is no single offence, an essential part of the defence consists in debating which provision really applies —or whether none does— with particular attention to the disputed application of Art. 197.7 CP to synthetic content.

None of these lines is generic. Each depends on the specific facts: the tool that generated the montage, how it was distributed, whether the victim is identifiable and, above all, their apparent age. A defence built on the disputed application of Art. 197.7 CP to synthetic content will not transfer to a case classified under Art. 189 CP, where the framework is far more severe and the room for manoeuvre much narrower. That is why an early and rigorous analysis of the classification, together with technical scrutiny of the evidence, is decisive from the very first steps of the proceedings.

For the victim, by contrast, the aim is to articulate the most solid classification, preserve the digital evidence as early as possible and pursue removal of the content through the DSA channels, in parallel with the criminal action.

Contact the Firm

Alonso Sala is a firm dedicated exclusively to criminal law, based at calle Velázquez 27 in Madrid and providing coverage throughout Spain. We analyze the classification of the facts among the various offences in play —Arts. 173, 197.7, 208 and 189 CP—, the strength of the computer forensic evidence and the most suitable strategy, whether in the defence of the accused or in the victim's private prosecution, always with technical rigor and the utmost discretion.