United States - Minnesota - Deepfake Liability (HF 1370)

Overview

Minnesota House File 1370 (HF 1370), enacted in May 2023 and effective August 1, 2023, represents one of the most comprehensive state-level approaches to regulating deepfake technology in the United States. The legislation creates both criminal penalties and civil causes of action for two primary categories of deepfake misuse: nonconsensual dissemination of deepfake sexual images and the use of deepfake technology to influence elections. Minnesota became one of the first states to specifically address AI-generated synthetic media in the electoral context with criminal penalties (Minnesota Revisor of Statutes).

The law was passed as part of a broader omnibus judiciary bill (Laws 2023, Chapter 52, Article 16, Section 18) and reflects growing national concern about the potential for AI-generated content to cause harm to individuals and democratic institutions. Minnesota joins Texas, California, and several other states in enacting legislation specifically targeting deepfake content, though Minnesota's approach is notable for addressing both intimate imagery and election integrity in a single statutory framework.

Definitions

The law provides precise definitions critical to its enforcement (Minn. Stat. § 609.771, Subd. 1):

• Deep fake: Any video recording, motion-picture film, sound recording, electronic image, or photograph, or any technological representation of speech or conduct substantially produced by artificial intelligence, machine learning, generative adversarial network techniques, or other computer-generated or machine-learning means that is so realistic that a reasonable person would believe it depicts speech by or conduct of an individual who did not in fact engage in such speech or conduct.
• Intimate parts: The primary genital area, groin, inner thigh, buttocks, or breast of a human being.
• Sexual act: Any act which involves sexual penetration, the touching of intimate parts in a sexual manner, or any lewd exhibition of the intimate parts.
• Depicted individual: The individual whose likeness, voice, or other identifying characteristics are used to create or alter the deep fake content.

Key Focus Areas

• Nonconsensual Sexual Deepfakes (Subd. 2): The law criminalizes the intentional dissemination of a deep fake that depicts the depicted individual with the intimate parts of the depicted individual exposed, or appearing to be exposed, when the actor knows or reasonably should know that the depicted individual did not consent. This includes deepfakes depicting individuals in sexual acts they never performed. The provision builds on Minnesota's existing revenge porn statute (Minn. Stat. § 617.261) but extends protection specifically to AI-generated content.
• Election-Related Deepfakes (Subd. 3): The law creates a separate, more serious offense for disseminating a deep fake of a candidate within 90 days of an election in which the candidate is running, with the intent to injure the candidate's reputation or to influence the result of the election. This provision applies to deepfakes depicting speech or conduct the candidate did not actually engage in and represents one of the strongest state-level protections against AI-generated election misinformation.
• Civil Remedies (Subd. 4): Beyond criminal penalties, the law provides victims with a private cause of action. Depicted individuals may sue for injunctive relief, actual damages, a civil penalty of up to $100,000, and reasonable attorney fees and costs. Courts may also issue temporary restraining orders and preliminary injunctions to prevent further dissemination.
• Platform Accountability: While the law preserves Section 230 protections for online platforms, it creates incentives for rapid takedown by establishing clear liability for original disseminators and potential civil exposure for those who continue to distribute after notification.

Governance and Institutional Framework

Enforcement of Minnesota HF 1370 operates through multiple channels within the state's existing legal infrastructure:

• Criminal Prosecution: County attorneys have authority to prosecute violations as gross misdemeanors or felonies depending on the circumstances. The Minnesota Bureau of Criminal Apprehension (BCA) may assist with investigations involving sophisticated digital forensics.
• Civil Enforcement: Minnesota district courts have jurisdiction over civil claims brought by depicted individuals. The law provides for expedited proceedings where injunctive relief is sought to prevent ongoing harm.
• Attorney General's Office: The Minnesota Attorney General may bring actions under the state's consumer protection statutes if deepfake activity constitutes unfair or deceptive practices.
• Coordination with Federal Law: The statute explicitly preserves federal immunity under 47 U.S.C. § 230 (Communications Decency Act) for interactive computer services, ensuring that online platforms are not held liable as publishers of third-party deepfake content unless they cross into active participation.

Implementation Framework

The law became effective on August 1, 2023, with immediate applicability to conduct occurring on or after that date. Key implementation considerations include:

• Prosecutorial Discretion: County attorneys retain discretion in charging decisions, weighing factors such as the extent of dissemination, harm to victims, and intent of the actor.
• Evidence Standards: Prosecutors must establish that content qualifies as a "deep fake" under the statutory definition, which may require expert testimony on AI-generation techniques and the realistic nature of the content.
• Victim Services: The law's civil remedies complement Minnesota's existing crime victim services, allowing depicted individuals to pursue both criminal justice and civil compensation simultaneously.

Monitoring and Evaluation

As a relatively recent law, formal monitoring mechanisms are still developing. Key aspects include:

• Judicial Reporting: Minnesota courts track cases filed under specific statutes, enabling assessment of how frequently the law is invoked.
• Legislative Oversight: The Minnesota House Judiciary Committee and Senate Judiciary and Public Safety Committee maintain ongoing oversight of criminal justice legislation.
• Law Enforcement Training: The Minnesota POST Board (Peace Officer Standards and Training) may incorporate deepfake identification into law enforcement training curricula.

Penalties, Liability, and Appeals

The law establishes a graduated penalty structure based on offense severity (Minn. Stat. § 609.771):

• Nonconsensual Sexual Deepfakes (First Offense): Gross misdemeanor punishable by up to 1 year imprisonment and/or a fine up to $3,000.
• Nonconsensual Sexual Deepfakes (Repeat Offense or Aggravated): Felony punishable by up to 3 years imprisonment and/or a fine up to $5,000. Aggravating factors include prior convictions, distribution to minors, or commercial exploitation.
• Election-Related Deepfakes: Felony punishable by up to 5 years imprisonment and/or a fine up to $10,000. The enhanced penalty reflects the public interest in election integrity.
• Civil Damages: Depicted individuals may recover actual damages (including emotional distress), statutory civil penalties up to $100,000, punitive damages in cases of willful and malicious conduct, and reasonable attorney fees and court costs.
• Injunctive Relief: Courts may order removal of deepfake content, prohibition on further dissemination, and disclosure of distribution channels to enable comprehensive takedowns.

Exceptions and Defenses

The law includes important exceptions to avoid chilling legitimate speech and activities (Minn. Stat. § 609.771, Subd. 5):

• Law Enforcement: Actions by law enforcement officers in the course of official duties are exempt.
• Legal Proceedings: Use of deepfake content as evidence in judicial, legislative, or administrative proceedings is permitted.
• Satire and Parody: Content clearly identified as satire, parody, or commentary that no reasonable person would believe to be authentic is protected.
• Educational and Research Purposes: Use for legitimate educational, documentary, or scientific research purposes is permitted when the content is clearly labeled as a deepfake.
• Consent: Dissemination with the knowing and voluntary consent of the depicted individual is not prohibited, though consent to creation does not automatically imply consent to distribution.

Relationship to Other Instruments

Minnesota HF 1370 exists within a complex framework of related state and federal laws:

• Minnesota Revenge Porn Statute (Minn. Stat. § 617.261): HF 1370 complements the existing revenge porn law by extending protection to AI-generated content. The two laws may apply in parallel where actual and synthetic intimate images are disseminated together.
• Federal Communications Decency Act (47 U.S.C. § 230): The law explicitly preserves Section 230 immunity for online platforms, ensuring the statute does not conflict with federal law.
• First Amendment: The law's exceptions for satire, parody, and political commentary reflect careful attention to First Amendment concerns. Courts will likely apply strict scrutiny to any enforcement that implicates protected speech.
• Other State Deepfake Laws: Minnesota joins Texas (2019), California (2019), and several other states in enacting deepfake-specific legislation. Minnesota's approach is notable for combining sexual imagery and election provisions in a single comprehensive statute.

International Alignment

While HF 1370 is state-level legislation, it reflects broader international trends in deepfake regulation:

• European Union AI Act: The EU's comprehensive AI legislation includes transparency requirements for deepfakes, requiring clear disclosure when AI-generated content depicts real persons. Minnesota's approach differs by focusing on prohibition and penalty rather than disclosure requirements.
• United Kingdom Online Safety Act: The UK's framework addresses deepfake intimate imagery through platform duty of care requirements. Minnesota's direct criminalization represents a more aggressive regulatory approach.
• International Standards: The ISO/IEC is developing standards for AI-generated content authentication and detection, which may inform future amendments to laws like HF 1370.

Implementation Timeline

Sources and References