United States - New Hampshire - Deepfakes in Elections (HB 1596)

Overview

New Hampshire House Bill 1596, now enacted as Chapter 345 of the 2024 Session Laws, represents a pivotal legislative effort to address the growing threat of artificial intelligence (AI)-generated misinformation in political campaigns. Titled "AN ACT requiring a disclosure of deceptive artificial intelligence usage in political advertising," this legislation was signed into law by Governor Chris Sununu on August 2, 2024, with an effective date of August 1, 2024. The primary objective of the Act is to safeguard the integrity of the electoral process by imposing strict regulations on the creation and dissemination of synthetic media and deepfakes that could mislead voters. This proactive measure by the New Hampshire General Court underscores a broader national and international concern regarding the potential for AI technologies to manipulate public perception and undermine democratic institutions, particularly during critical election periods. The law specifically targets instances where AI is used to create realistic but false depictions of candidates or election officials, aiming to prevent the spread of materially deceptive content that could sway election outcomes.

The impetus for this legislation stems from a recognition of the rapid advancements in AI technology, which have made it increasingly easy to generate highly convincing fake audio, video, and images. These "deepfakes" pose a significant challenge to truth and trust in political discourse, as they can be used to fabricate statements or actions by individuals that never occurred, thereby creating confusion and eroding public confidence. The New Hampshire Act seeks to mitigate this risk by establishing clear rules for transparency and accountability. It introduces a new section, RSA 664:14-c, into the state's election laws, which outlines prohibitions and disclosure requirements for the use of such media in political advertising. By focusing on the deceptive nature of these AI-generated materials and their potential impact on elections, the Act aims to strike a balance between allowing for legitimate forms of political expression and preventing malicious manipulation, thereby protecting both candidates and the electorate from sophisticated forms of digital deception.

Definitions

The Act establishes precise definitions to delineate the scope of its prohibitions and requirements, ensuring clarity in its application to emerging AI technologies. Central to the legislation are the terms "synthetic media" and "deepfake." "Synthetic media" is defined broadly as an image, an audio recording, or a video recording of an individual's appearance, speech, or conduct that has been created or manipulated using generative adversarial network techniques or other digital technologies to produce a realistic but false depiction. This definition captures a wide array of AI-generated content, emphasizing the artificial creation or alteration of media to present something that did not genuinely occur. The inclusion of "generative adversarial network techniques" highlights the specific AI methods often employed in creating highly convincing fabricated content, while also allowing for the inclusion of other digital manipulation technologies.

Building upon the definition of synthetic media, the Act further defines "deepfake" as a specific type of media where a person's face, body, or voice has been digitally altered so that he or she appears to be someone else, appears to be saying something that he or she has never said, or appears to be doing something that he or she has never done. This definition underscores the deceptive intent and potential for misrepresentation inherent in deepfake technology. The legislation also provides definitions for "Artificial Intelligence (AI)," describing it as the capability of a machine to exhibit human-like cognitive functions and adapt its behavior without significant human oversight, and "Generative AI," which is AI capable of generating text, images, or other media in response to prompts. These foundational definitions are crucial for enforcing the Act, particularly when determining whether a piece of political advertising constitutes a "deceptive and fraudulent deepfake" – synthetic media that provides a fundamentally inaccurate impression and is intended to mislead the public about a depicted individual's actions or statements. The Act also implicitly defines "sponsor" as the entity responsible for distributing such media, and "candidate" or "election official" as the individuals whose likeness or voice is being manipulated, thereby establishing who is protected and who is liable under the law.

Governance and Institutional Framework

The enforcement and oversight of New Hampshire's deepfake legislation are primarily vested within existing state governmental structures, leveraging established mechanisms for election law compliance. The Department of Justice's Election Law Unit is explicitly designated as the body responsible for investigating and prosecuting reported violations of this Act. This allocation of responsibility ensures that the enforcement of regulations concerning deceptive AI usage in political advertising is integrated into the broader framework of election law enforcement, benefiting from the unit's specialized expertise and established procedures for handling electoral misconduct. The Election Law Unit's role is critical in maintaining the integrity of the state's elections by actively monitoring for and responding to instances where synthetic media or deepfakes are used to mislead voters or unfairly target candidates and election officials. Their involvement underscores the state's commitment to treating AI-driven disinformation as a serious threat to democratic processes, requiring dedicated legal and investigative resources.

The Act's placement within RSA 664, which governs elections, further solidifies its institutional framework by embedding it directly into the state's comprehensive election code. This integration means that the new provisions regarding deepfakes are subject to the same legal interpretations, procedural rules, and enforcement powers as other election offenses. While the Act does not establish a new regulatory body, it empowers existing authorities to adapt to the challenges posed by advanced digital manipulation. The process for addressing violations typically involves complaints being filed with the Department of Justice, which then initiates investigations. The courts play a crucial role in providing injunctive relief and awarding damages, as outlined in the Act, thereby serving as the ultimate arbiter in disputes arising from deceptive deepfake usage. This framework relies on a combination of administrative investigation and judicial review to ensure accountability and provide remedies for those harmed by violations, reflecting a pragmatic approach to regulating rapidly evolving technology within established legal systems.

Key Focus Areas

The New Hampshire Act on deepfakes in elections centers on several critical areas to combat the deceptive use of AI in political discourse. A primary focus is the outright prohibition of distributing deceptive and fraudulent deepfakes of candidates or parties on state or local ballots within 90 days of an election. This 90-day window is a strategic period leading up to an election when voters are most engaged and susceptible to last-minute misinformation, and when there is often insufficient time for targeted individuals to effectively debunk false narratives. The prohibition applies when the person or entity distributing the synthetic media "knows or should have known" it constitutes a deceptive deepfake, establishing a standard of reasonable diligence and accountability. This provision is designed to prevent malicious actors from intentionally spreading fabricated content that could materially affect an election's outcome by misrepresenting a candidate's or election official's appearance, actions, or speech.

Beyond outright prohibition, the Act also mandates clear and conspicuous disclosure requirements for any synthetic media used in political advertising that falls outside the prohibited category or is otherwise permissible. For visual media, the disclosure text must be easily readable by the average viewer, no smaller than the largest font size of other text, and, in the case of video, must appear for the duration of the video. For audio-only media, a clear statement identifying the content as AI-generated must be included at the beginning and end. This emphasis on transparency is crucial, as it allows the public to be aware when they are consuming manipulated content, thereby fostering media literacy and critical thinking. The Act also carves out specific exceptions to these rules, notably for satire, parody, or productions substantially dependent on physical or verbal impersonation without reliance on artificial intelligence. These exceptions are vital for protecting legitimate forms of artistic expression and political commentary, ensuring that the law does not inadvertently stifle free speech while still addressing the core problem of deceptive AI-generated misinformation. The scope of the Act extends to protect not only candidates but also election officials, acknowledging their critical role in the democratic process and their vulnerability to similar forms of digital manipulation.

Implementation Framework

The implementation framework for New Hampshire's deepfake legislation is designed to provide clear guidelines for compliance and enforcement, focusing on the critical period leading up to an election. A central element of this framework is the establishment of a 90-day pre-election prohibition period. Within this timeframe, it is explicitly forbidden for any person, corporation, committee, or other entity to distribute synthetic media that they know or should have known is a deceptive and fraudulent deepfake of a candidate or party on the state or local ballot. This "knows or should have known" standard is significant, as it places a burden on distributors to exercise due diligence in verifying the authenticity of political advertising content, rather than solely relying on actual malicious intent. It implies that ignorance or willful blindness will not serve as a defense against violations, encouraging responsible dissemination practices, particularly in the highly sensitive electoral context. The Act's focus on distributors, rather than just creators, broadens the scope of accountability, recognizing that the act of sharing deceptive content can be as damaging as its initial creation.

Furthermore, the implementation framework details specific requirements for disclosure when synthetic media is used in political advertising outside the scope of the outright prohibition. For visual media, the disclosure must be prominently displayed, with text size and duration requirements designed to ensure it is easily noticeable and readable by the average viewer. In video content, this disclosure must remain visible for the entire duration of the video. For audio-only messages, a distinct disclosure must be provided at both the beginning and end of the communication. These explicit formatting rules aim to prevent disclosures from being hidden or easily missed, thereby maximizing their effectiveness in informing the public. The Act also outlines the types of media covered, including images, audio recordings, and video recordings, ensuring comprehensive coverage across various digital formats. By setting clear standards for both prohibition and disclosure, the implementation framework seeks to create a predictable legal environment that discourages deceptive practices while allowing for legitimate, transparent uses of AI in political messaging, ultimately fostering a more informed and resilient electorate against digital manipulation.

Monitoring and Evaluation

While New Hampshire HB 1596 does not explicitly detail a formal monitoring and evaluation framework with specific metrics or reporting requirements, the ongoing enforcement role of the Department of Justice's Election Law Unit inherently involves a continuous process of monitoring and assessment. The unit, responsible for upholding the state's election laws, will naturally track complaints, investigations, and prosecutions related to deceptive deepfakes. This operational monitoring will provide practical insights into the prevalence of AI-generated misinformation in New Hampshire elections, the types of deepfakes being used, and the effectiveness of the Act's provisions in deterring such practices. The frequency and nature of reported violations, the outcomes of legal actions (e.g., injunctions granted, damages awarded), and the challenges encountered during investigations will serve as de facto indicators of the law's impact and areas where further legislative or enforcement adjustments might be necessary. The absence of a dedicated evaluation committee or mandated annual reports suggests that the assessment of the Act's efficacy will be integrated into the routine functions and operational reviews of the Election Law Unit, rather than through a separate, formalized process.

The broader legislative context also contributes to an informal evaluation process. As a pioneering piece of state legislation addressing AI deepfakes in elections, its implementation will likely be observed by other states and national bodies grappling with similar challenges. The experiences and outcomes in New Hampshire could inform future legislative debates and policy developments elsewhere. Furthermore, the legal remedies provided by the Act, such as the ability for candidates and election officials to seek injunctive relief and damages, will generate a body of case law that can be analyzed to understand how courts interpret and apply the law. This judicial interpretation will further refine the practical understanding of the Act's provisions, including the "knows or should have known" standard and the definition of "deceptive and fraudulent deepfake." While not a formal evaluation mechanism, the aggregation of these legal precedents and enforcement experiences will collectively contribute to an ongoing assessment of the Act's strengths, weaknesses, and overall effectiveness in achieving its goal of protecting election integrity from AI-driven deception.

Penalties, Liability, and Appeals

New Hampshire HB 1596 establishes clear legal consequences for violations, outlining both the types of remedies available and the parties who can be held liable. The Act empowers a candidate or election official whose appearance, action, or speech is depicted through the use of a deceptive and fraudulent deepfake, in violation of the established prohibitions, to seek injunctive or other equitable relief. This provision is critical because it allows for immediate legal action to prevent the further publication or distribution of harmful deepfakes, which can be essential in the fast-paced environment of an election campaign. Injunctive relief aims to stop the spread of misinformation before it can cause irreparable damage to a candidate's reputation or mislead a significant portion of the electorate. The ability to obtain such relief underscores the urgency recognized by the legislature in addressing this form of digital manipulation.

In addition to injunctive relief, the Act also permits a candidate or election official who has been targeted by a deceptive deepfake to bring an action for general or special damages against the sponsor responsible for its distribution. This allows for financial compensation for harm suffered, which could include reputational damage, campaign costs incurred to counter the deepfake, or other measurable losses. Furthermore, the court has the discretion to award a prevailing party reasonable attorneys' fees and costs, which incentivizes individuals to pursue legal action against violators and helps to offset the financial burden of litigation. Liability primarily falls on the "sponsor"—the person, corporation, committee, or other entity that distributes the deceptive synthetic media. The standard for liability is based on whether the sponsor "knows or should have known" that the distributed media was a deceptive and fraudulent deepfake, placing a responsibility on those disseminating political advertising to verify its authenticity. While the Act does not explicitly detail an appeals process, any judgments or orders issued by the courts under this section would be subject to the standard appellate procedures within the New Hampshire judicial system, allowing for review by higher courts.

Relationship to Other Instruments

New Hampshire HB 1596 is specifically designed to amend an existing legal framework, integrating its provisions directly into the state's established election laws. The Act inserts a new section, RSA 664:14-c, into Chapter 664 of the New Hampshire Revised Statutes Annotated, which governs elections. This integration means that the regulations concerning deceptive AI usage in political advertising become an intrinsic part of the comprehensive legal structure that dictates how elections are conducted in the state. By amending existing statutes, the legislature ensures consistency with established election principles and leverages the enforcement mechanisms already in place for other electoral offenses. This approach avoids creating a separate, potentially conflicting, regulatory regime and instead strengthens the existing legal framework to address modern technological challenges. The bill's relationship with RSA 664 ensures that its definitions, prohibitions, and remedies are understood within the broader context of campaign finance, political advertising, and election integrity regulations.

The Act's provisions also interact with fundamental constitutional principles, particularly those related to free speech under the First Amendment of the United States Constitution and similar protections in the New Hampshire State Constitution. While the legislation aims to prevent deceptive practices, concerns were raised during its legislative process about potential infringements on free speech. The inclusion of explicit exceptions for satire, parody, and physical impersonation without AI demonstrates an attempt by lawmakers to balance the need to combat misinformation with the protection of legitimate forms of political expression and artistic commentary. This balancing act is crucial in the legal landscape, as any restrictions on speech, even deceptive speech, must be carefully tailored to meet strict constitutional scrutiny. The Act's focus on "deceptive and fraudulent" deepfakes, particularly those intended to mislead, aligns with legal precedents that generally afford less protection to speech that is false and misleading, especially in commercial or electoral contexts. Furthermore, the Act complements other recent legislative efforts in New Hampshire addressing AI, such as HB 1432-FN (prohibiting fraudulent use of deepfakes with criminal penalties) and HB 1688-FN (regulating AI use by state agencies), indicating a coordinated approach to AI governance across different sectors.

International Alignment

As a state-level legislative act within the United States, New Hampshire HB 1596 does not directly aim for international alignment in its specific provisions or implementation. Its primary focus is on regulating political advertising within state and local elections in New Hampshire, addressing concerns specific to its jurisdiction. However, the Act's existence and its approach to regulating deepfakes in elections contribute to a broader global conversation and emerging trend among democratic nations to address the challenges posed by artificial intelligence in political discourse. Many countries and international bodies are grappling with similar issues of AI-generated misinformation, disinformation, and their potential impact on electoral integrity and public trust. Therefore, while not explicitly aligning with international standards, New Hampshire's legislation serves as a case study and a contribution to the global body of regulatory responses to AI's societal implications.

The principles underpinning HB 1596, such as transparency through disclosure requirements and accountability for deceptive content, resonate with discussions and recommendations from international organizations like the OECD, UNESCO, and the European Union, which are developing frameworks for trustworthy AI and combating disinformation. For instance, the EU's AI Act and various national strategies often emphasize the need for transparency in AI systems, especially those that generate content or influence public opinion. The New Hampshire Act's focus on defining "synthetic media" and "deepfakes" and establishing clear prohibitions and remedies reflects a common regulatory strategy seen globally: identifying the harmful use cases of AI and implementing targeted legal interventions. Thus, while its scope is strictly domestic, the New Hampshire deepfake law is part of a growing global mosaic of regulatory efforts, demonstrating a shared recognition of the urgent need to govern AI's impact on democratic processes and information environments.

Implementation Timeline

Compliance Checklist

Sources and References