Minnesota Adopts the First US Ban on “Nudification” Apps
On 7 May 2026, Minnesota became the first US state to prohibit access to so-called “nudification” applications. The law will enter into force on 1 August 2026.
These services use a photograph of an identifiable individual to generate a realistic image or video artificially depicting that person’s intimate parts.
Requiring no particular technical expertise, such tools enable the large-scale production of technology-facilitated sexual abuse. Since 2023, 23 cases involving school communities in the United States have reportedly already been identified.
A Preventive Ban Targeting the Tool Itself
The legislation does not solely target users who create or disseminate the content. It prohibits any person controlling a website, application, software program or service from:
- enabling access to, the download of or the use of a nudification function;
- generating the image on behalf of a user; or
- promoting such services.
Victims may be awarded up to three times the amount of their actual loss, as well as punitive damages, injunctive relief and legal costs.
The Minnesota Attorney General may also seek a civil penalty of up to USD 500,000 for each unlawful access, download or use.
Comparison with the European Union’s Recent Position
By way of reminder, on 16 June 2026, the European Parliament approved the inclusion in the AI Act of a prohibition targeting systems that generate child sexual abuse material or depict, without consent, the intimate parts or sexual activities of an identifiable individual.
The proposed framework applies to:
- providers placing such systems on the European market;
- systems that lack adequate technical safeguards; and
- deployers using them for such purposes.
Compliance would be required by 2 December 2026. Formal adoption by the Council remains necessary before the measure can enter into force.
Minnesota prohibits access to a specialised service, subject to an exception for tools requiring genuine technical expertise. The European Union has adopted a more systemic approach: it intervenes at the market-placement stage and requires safeguards to be incorporated by design.
In both cases, liability no longer rests solely with the person who creates the deepfake. It is shifting towards those who design, distribute or otherwise enable access to the tool.
This is one of the issues addressed by FIRSH in its recent work and hearings, including its hearing before the French High Council for Equality between Women and Men: should the law merely sanction the content, or should it also prevent its creation through system design?
Public authorities are gradually recognising, in line notably with the recommendations set out in our White Paper, that effective AI regulation requires action not only against harmful outcomes, but also on the architecture of the systems that make those outcomes foreseeable and reproducible.
