Deepfakes and the Protection of Personality Rights: Legal Remedies Against Synthetic Media

By 2026, deepfake technology has reached a level of sophistication that fundamentally undermines the credibility of visual and auditory evidence. Artificial intelligence can generate photorealistic images, videos, and audio recordings within seconds, depicting a real person in a situation or context in which they never participated. This technological reality poses challenges not only to the media and public trust but also opens new dimensions for the protection of personality rights. The following analysis examines the Hungarian civil law, criminal law, and EU regulatory frameworks.

Civil Law Protection: Likeness, Biometric Identity, and Non-Pecuniary Damages

The Right to One’s Own Image and Voice Recording

The Hungarian Civil Code (Polgári Törvénykönyv, hereinafter: Ptk.) provides for the general protection of personality rights in Section 2:42, and for specific personality rights in Section 2:43. Of paramount importance for deepfakes is Section 2:48 of the Ptk., which protects the right to one’s own image and voice recording:

“The consent of the person concerned shall be required for the making and use of an image or voice recording of that person.”

The 2026 judicial practice interprets this protection expansively: it sanctions not only the unauthorised use of actual recordings, but also the creation and dissemination of AI-generated synthetic “likenesses”. In the courts’ assessment, deepfake content constitutes an infringement of the person’s biometric and visual identity, irrespective of whether the recording is technically “authentic” or synthetic.

The Evolution of Non-Pecuniary Damages

The amount of non-pecuniary damages (sérelemdíj) under Section 2:52 of the Ptk. has increased significantly by 2026, reflecting the particular gravity of harm in deepfake cases. The courts weigh the following factors:

• The nature of the infringement: Cases involving intimate content produced without consent (intimate image-based abuse) or politically motivated defamation attract elevated damages.
• The extent of dissemination: Mass sharing on social media substantially increases the damages awarded.
• The perpetrator’s intent: Deliberate intent to cause harm (e.g. extortion, revenge, defamation of competitors) is treated as an aggravating factor.
• The victim’s vulnerability: Enhanced protection is afforded where minors, public figures, or particularly vulnerable individuals are concerned.

Important: The claim for non-pecuniary damages is based on a system of objective liability [Section 2:52(2) Ptk.] — establishing the fact of infringement does not require proof of the infringer’s fault, though the court may take culpability into account when determining the amount of damages.

Criminal Law Implications

Applicable Offences Under the Criminal Code

Several offences under Act C of 2012 on the Criminal Code (hereinafter: Btk.) may apply to deepfake-related conduct:

1. Misuse of personal data (Section 219 Btk.): Where the deepfake content involves the unauthorised processing of the victim’s personal data (facial likeness, biometric data).
2. Defamation (Section 226 Btk.): Where the deepfake content conveys a statement of fact capable of damaging the victim’s reputation — for example, depicting the victim in a fabricated situation.
3. Insult (Section 227 Btk.): Where the content violates the victim’s human dignity before the public.
4. Harassment (Section 222 Btk.): Where deepfake content is used as a tool for systematic intimidation or psychological abuse.

Legislative Developments in 2026

In response to the proliferation of deepfake technology, the legislature has in 2026 introduced a supplementary offence in the Btk. to protect electoral processes and democratic institutions. Under these new provisions:

• The production and dissemination of intentionally misleading deepfake content during an election campaign constitutes an independent criminal offence.
• A qualified case applies where the content is capable of influencing voters’ will or gravely undermining public trust in democratic institutions.
• The penalty depends on the gravity of the conduct and may be classified as a misdemeanour or a felony.

Note: Applying traditional criminal offences (defamation, insult) to deepfake cases is not always straightforward, as the technological context raises novel evidentiary questions — particularly regarding intent and causation.

The AI Act’s Deepfake-Related Provisions

Article 50 of Regulation (EU) 2024/1689 (the AI Act) imposes express transparency obligations for deepfake content:

• Image, audio, or video content produced using deepfake technology must be clearly labelled as AI-generated.
• The labelling must include a machine-readable watermark.
• The obligation applies to both the producer and the disseminator of the content.

Under Article 99 of the AI Act, failure to comply with the labelling obligation may result in administrative fines of up to 3% of global annual turnover.

Practical nexus: The AI Act’s labelling obligation and the Ptk.’s personality rights protection form a mutually reinforcing system: the absence of a watermark not only triggers an administrative fine but may also establish the infringer’s bad faith in personality rights proceedings.

Platform Liability Under the DSA

Regulation (EU) 2022/2065 on digital services (hereinafter: the DSA — Digital Services Act) has established a comprehensive liability framework for intermediary service providers:

Notice and Removal Obligations

Under Article 16 of the DSA, platforms must provide an effective notice mechanism enabling any person to report illegal content. Under Article 22, reports from trusted flaggers must be treated with priority.

In the case of deepfake content, platforms are required to:

1. Promptly initiate an investigation upon receipt of a notice.
2. Remove content that is manifestly unlawful (e.g. intimate deepfake content, election manipulation).
3. Notify the user who produced the content of the reason for removal and available remedies.

Automated Detection

By 2026, major platforms employ automated AI-based detection algorithms to identify deepfake content. Under Article 34 of the DSA, Very Large Online Platforms (VLOPs) are required to conduct regular risk assessments covering the systemic risks posed by deepfake technology. Automated systems do not, however, relieve platforms of the obligation to conduct individual assessments.

Evidentiary Challenges in Deepfake Cases

One of the most complex challenges in deepfake cases is the handling of digital evidence:

• Forensic expert analysis: Courts increasingly appoint AI-technology experts to determine whether a recording is authentic or synthetically generated.
• Metadata analysis: The metadata of recordings (EXIF data, digital watermarks, provenance information) have become a key evidentiary tool.
• Blockchain-based content provenance: Certain organisations and media outlets have in 2026 adopted blockchain technology to verify the authenticity of content (content provenance).

Practical Recommendations for Victims

If you or a family member has become a victim of deepfake content, we recommend the following steps:

1. Preserve evidence: Take screenshots and save the content’s URL. The digital footprint may disappear quickly once the content is removed.
2. Report to the platform: Use the platform’s notice mechanism, referencing Article 16 of the DSA and the unlawful nature of the content.
3. Seek legal advice: Contact a lawyer to pursue interim injunctive relief and civil claims without delay.
4. File a criminal report: In serious cases (intimate content, extortion, electoral manipulation), file a report with the police.
5. Authority referral: The National Authority for Data Protection and Freedom of Information (NAIH) may also take action in cases involving the misuse of personal data.

Conclusion

Deepfake technology presents the system of personality rights protection with unprecedented challenges, to which the Hungarian legal order in 2026 provides a multi-layered response: the Ptk.’s personality rights protection, the Btk.’s criminal sanctions, the AI Act’s transparency obligations, and the DSA’s platform liability framework collectively ensure legal redress. The most effective course of action is founded on immediate evidence preservation, platform reporting, and the enforcement of legal claims.

If you have become a victim of deepfake content, contact our office — our expert team will assist you in enforcing your rights without delay.