United States - Hawaii - Deepfakes Regulation (SB 2687)

Overview

Hawaii Senate Bill 2687 (SB 2687), enacted as Act 191 on July 3, 2024, represents a significant legislative effort by the State of Hawaii to address the growing concerns surrounding the malicious use of artificial intelligence (AI) and deepfake technology, particularly within the electoral process. The legislature recognized the dual nature of AI, acknowledging its potential for societal benefit while simultaneously highlighting its dangerous consequences when misused, especially in spreading disinformation and misinformation during elections. This Act aims to safeguard the integrity of democratic processes by directly confronting the threat posed by AI-generated deepfakes that can manipulate public perception and voter behavior. The core objective of SB 2687 is multifaceted: it prohibits the reckless distribution of 'materially deceptive media' during critical election periods, establishes clear remedies for individuals and organizations harmed by such distribution, and institutes criminal penalties for violations. By doing so, Hawaii joins a growing number of states, including Michigan, Minnesota, and Washington, that have enacted legislation to regulate AI in elections, underscoring a nationwide recognition of the vulnerabilities that advanced AI technologies introduce to political discourse and electoral fairness.

The scope of Act 191 is specifically tailored to election cycles, applying its prohibitions from the first working day of February in every even-numbered year through the next general election. This targeted approach reflects the legislature's intent to protect the most sensitive periods of political campaigning and voter decision-making from the undue influence of fabricated content. The law defines 'materially deceptive media' broadly to include AI-generated or substantially altered video, audio, or images that falsely depict an individual engaging in speech or conduct they did not, in a manner convincing to a reasonable person. Crucially, the prohibition is triggered when such media is distributed with reckless disregard for harming a candidate's reputation or electoral prospects, or with the intent to alter voting behavior by deceiving electors. The Act also outlines specific exceptions, primarily for legitimate news reporting, satire, or parody, provided a clear and conspicuous disclaimer is present. These provisions collectively form a robust framework designed to mitigate the risks of AI-driven electoral interference while attempting to balance free speech considerations.

Definitions

Central to the enforcement and understanding of Hawaii Act 191 are its precisely defined terms, which delineate the scope of prohibited activities and the nature of the media it seeks to regulate. The Act introduces and clarifies several key concepts, paramount among them being 'materially deceptive media.' This term refers to any video, audio, or image content that has been generated or substantially altered through the use of artificial intelligence, and which falsely portrays a 'depicted individual' engaging in speech or conduct that they did not, in fact, undertake. The critical element of this definition is that the false portrayal must be convincing to a 'reasonable person,' establishing an objective standard for assessing the deceptive quality of the media. Furthermore, for media to fall under the Act's prohibition, its distribution must be undertaken with reckless disregard for harming a candidate's reputation or electoral prospects, or with the intent to change voter behavior by misleading them into believing the depicted individual actually engaged in the shown or said actions. This focus on both the technical nature of the content and the intent behind its distribution is crucial for distinguishing harmful deepfakes from legitimate forms of expression.

Another fundamental definition is that of 'deepfake' and 'generative AI,' which, while not explicitly defined as standalone terms within the operative sections of the Act, are implicitly understood through the description of 'materially deceptive media' and the legislative findings. The legislature explicitly states its concern with "the use of deepfakes or generative AI in elections" as powerful tools for spreading disinformation and misinformation. This contextual understanding highlights that the Act targets the output of these advanced AI technologies when they are used to create realistic but fabricated content. The 'depicted individual' is also a crucial concept, referring to the person whose likeness, voice, or actions are falsely represented in the materially deceptive media. This individual, whether a candidate for election or any other person, is granted specific remedies under the Act if their appearance, action, or speech is altered or affected through such media. The 'election period' is also narrowly defined as the timeframe from the first working day of February in every even-numbered year through the next general election, ensuring the law's applicability during the most critical phases of political campaigns. These definitions collectively provide the legal framework necessary for identifying, prosecuting, and seeking redress for the harmful impacts of AI-generated deceptive content in the electoral sphere.

Governance and Institutional Framework

The governance and institutional framework for enforcing Hawaii Act 191 primarily relies on existing legal structures, empowering various state authorities to address violations related to materially deceptive media. While the Act does not establish a new dedicated regulatory body for deepfakes, it integrates the regulation of such content into the purview of established legal and electoral enforcement mechanisms. The primary responsibility for prosecuting criminal violations falls upon the Attorney General, county attorneys, and county prosecutors, who are tasked with investigating and bringing charges against individuals or entities that recklessly distribute materially deceptive media during election periods. This approach leverages the existing prosecutorial infrastructure to ensure that harmful deepfakes are treated with the seriousness of other electoral offenses. The Act also grants civil remedies, allowing injured parties to bring direct actions in court, further decentralizing enforcement and providing avenues for private redress.

Beyond criminal prosecution, the Hawaii Campaign Spending Commission may also play a role in the enforcement of this Act. The legislation specifies that the Commission has the authority to assess fines for violations or refer violations for criminal prosecution, particularly when the deceptive media pertains to electoral campaigns and potentially impacts campaign finance or fair election practices. This involvement of the Campaign Spending Commission highlights the Act's focus on maintaining integrity within the electoral process and ensuring that AI-generated disinformation does not undermine the fairness of elections. Furthermore, the courts are central to the enforcement framework, as they are empowered to issue temporary or permanent injunctions or restraining orders to prevent further harm from the distribution of materially deceptive media. They can also award general or special damages, as well as reasonable attorney's fees and costs, to prevailing plaintiffs. This multi-pronged approach, involving both criminal and civil avenues, and engaging various levels of government and judicial authority, is designed to provide comprehensive oversight and enforcement against the malicious use of deepfakes in Hawaii's elections.

Key Focus Areas

Hawaii Act 191 zeroes in on several critical areas to combat the misuse of deepfake technology, with a primary emphasis on protecting the integrity of elections. The foremost focus is the prohibition on the distribution of materially deceptive media during defined election periods. This prohibition is not absolute but is specifically triggered when the distribution is made with reckless disregard for harming a candidate's reputation or electoral prospects, or with the intent to deceive voters into believing a depicted individual engaged in fabricated speech or conduct. This targeted approach acknowledges that not all AI-generated content is harmful, but rather focuses on content created and disseminated with malicious intent to influence election outcomes. The Act aims to prevent scenarios where voters are misled by realistic but false portrayals of candidates or other individuals, thereby safeguarding the democratic process from manipulation.

Another key focus area is transparency through disclaimer requirements. Recognizing that some AI-generated content may have legitimate uses (e.g., for satire, parody, or news reporting), the Act provides exceptions if the media includes a clear and conspicuous disclaimer. This disclaimer must explicitly state that the content has been altered or generated by artificial intelligence and must appear throughout the entire duration of the media, whether it's video, audio, or an image. This provision ensures that even when AI-generated content is permitted, the audience is informed of its artificial nature, allowing them to critically assess the information presented. This balances the need to combat disinformation with the protection of free speech, ensuring that creative or journalistic uses of AI are not unduly stifled, provided they are transparent. Furthermore, the Act establishes criminal penalties and civil remedies for violations, creating a strong deterrent against non-compliant behavior and providing recourse for those harmed. This includes penalties ranging from petty misdemeanors to class C felonies, depending on the severity and intent of the violation, as well as the ability for depicted individuals or voter interest organizations to seek damages and injunctive relief.

Implementation Framework

The implementation framework for Hawaii Act 191 is designed to integrate its provisions into existing legal and electoral structures, ensuring that the regulation of materially deceptive media is enforceable and effective without requiring the creation of entirely new administrative bodies. The Act's core prohibitions become active during specific "election periods," defined as the timeframe from the first working day of February in every even-numbered year through the next general election. This temporal specificity ensures that enforcement efforts are concentrated during the most critical phases of political campaigning, when the potential for harm from deepfakes is highest. The law places the onus on individuals and entities to exercise due diligence and refrain from recklessly distributing prohibited media, thereby fostering a sense of responsibility among those involved in political communication. Education and awareness campaigns, while not explicitly mandated by the bill, would likely be a crucial informal component of effective implementation, informing candidates, media outlets, and the public about the new regulations and the risks associated with deepfakes.

Furthermore, the implementation relies heavily on the judicial system for both criminal prosecution and civil redress. Law enforcement agencies, including the Attorney General and county prosecutors, are responsible for investigating alleged criminal violations and bringing appropriate charges. For civil actions, individuals whose likeness or voice is deceptively used, as well as organizations representing the interests of potentially deceived voters, are empowered to initiate legal proceedings. This allows for a decentralized enforcement mechanism where affected parties can directly seek justice. Courts are granted broad powers to issue injunctions, award damages, and impose civil fines for non-compliance with court orders, providing robust tools for preventing and remedying harm. The inclusion of the Hawaii Campaign Spending Commission in the enforcement framework further strengthens the implementation by allowing for administrative fines and referrals for criminal prosecution, particularly for violations that directly impact the fairness and integrity of elections. The Act's framework also includes specific exceptions for broadcasters, cable operators, and internet service providers, provided they do not know the content is being published with deceptive intent, which is a practical consideration for modern media distribution.

Monitoring and Evaluation

While Hawaii Act 191 does not explicitly detail a formal, standalone monitoring and evaluation framework, its effectiveness will inherently be assessed through the outcomes of its enforcement and the ongoing observation of its impact on the electoral landscape. The primary mechanism for monitoring the Act's efficacy will be the number and nature of cases brought under its provisions, both criminal prosecutions by state and county attorneys and civil actions initiated by injured parties. Tracking these cases will provide insights into the prevalence of materially deceptive media during election periods, the types of deepfakes being disseminated, and the success rates of enforcement actions. This data will be crucial for understanding whether the Act serves as an effective deterrent against malicious deepfake usage and whether the established penalties and remedies are sufficient to address the harm caused. The judicial system, through its rulings and the precedents it sets, will also play a significant role in shaping the interpretation and application of the Act, thereby influencing its long-term impact.

Beyond formal legal proceedings, the broader impact of Act 191 will be informally monitored by stakeholders such as election officials, political campaigns, media organizations, and civil society groups focused on election integrity and digital rights. These groups will likely observe changes in the landscape of political communication, noting any decrease in the spread of deepfakes or an increase in the use of disclaimers on AI-generated content. Public awareness and compliance with the disclaimer requirements will also be key indicators of the Act's success in promoting transparency. As deepfake technology continues to evolve rapidly, the Hawaii Legislature will likely engage in ongoing review of the Act's provisions. The legislative findings themselves suggest an understanding that existing laws may need revision as AI technology advances, implying a commitment to adapt the regulatory framework to emerging challenges. Future amendments or supplementary legislation may be considered based on practical experiences with the Act's implementation, technological advancements, and any unforeseen consequences, ensuring the law remains relevant and effective in protecting democratic processes.

Penalties, Liability, and Appeals

Hawaii Act 191 establishes a clear tiered system of penalties and liabilities for violations related to the distribution of materially deceptive media, reflecting the legislature's intent to deter malicious use of deepfakes and provide robust recourse for those harmed. For general violations, a person found guilty of recklessly distributing materially deceptive media is subject to a petty misdemeanor charge. This initial level of penalty serves as a foundational deterrent, signaling that such actions are not taken lightly by the state. However, the Act escalates penalties for repeat offenders, stipulating that a person who violates the section within five years of a previous conviction for the same offense shall be guilty of a misdemeanor. This provision aims to address habitual offenders and reinforce the seriousness of repeated transgressions in undermining electoral integrity.

The most severe criminal penalties are reserved for violations committed with the intent to cause violence or bodily harm, in which case a person shall be guilty of a class C felony. This recognizes the potential for deepfakes to incite real-world harm and violence, placing such malicious intent at the highest level of criminal culpability. In addition to criminal prosecution, the Act provides significant civil remedies. A 'depicted individual' whose appearance, action, or speech is altered or affected through materially deceptive media, or any organization representing the interests of voters likely to be deceived, may bring an action for general or special damages against the violator. Courts are empowered to issue temporary or permanent injunctions or restraining orders to prevent further harm, and may award reasonable attorney's fees and costs to a prevailing party. Furthermore, the Hawaii Campaign Spending Commission has the authority to assess civil fines for violations or refer cases for criminal prosecution. Courts can also issue civil fines of up to $1,000 per day for violations of court orders. The Act does not explicitly detail an appeals process, but appeals would follow standard judicial procedures for criminal convictions and civil judgments within the Hawaii state court system, allowing for review by higher courts.

Relationship to Other Instruments

Hawaii Act 191, while a standalone piece of legislation specifically targeting deepfakes in elections, operates within the broader legal and regulatory landscape of the State of Hawaii and the United States. It complements existing laws related to election integrity, defamation, fraud, and campaign finance. For instance, the Act's provisions on criminal penalties and civil remedies would interact with Hawaii's existing criminal code and civil procedure rules, dictating how prosecutions are conducted and how civil claims are litigated. The establishment of remedies for injured parties, including the ability to seek damages and injunctive relief, aligns with established tort law principles, providing a specific application of these principles to the unique harms posed by deepfakes. The involvement of the Hawaii Campaign Spending Commission in assessing fines or referring violations for criminal prosecution indicates a direct relationship with existing campaign finance regulations, ensuring that the integrity of political spending and communication is maintained in the face of new technological threats.

The Act also implicitly interacts with constitutional protections, particularly those related to freedom of speech. The inclusion of exceptions for satire, parody, and news reporting, along with the requirement for a clear disclaimer, demonstrates an attempt by the legislature to balance the need to regulate deceptive content with the constitutional right to free expression. This careful drafting aims to avoid challenges based on overbreadth or vagueness that could arise if the prohibitions were too broad. Furthermore, the legislative findings acknowledge that other states, such as Michigan, Minnesota, and Washington, have enacted similar legislation governing AI in elections. This indicates an awareness of a national trend in deepfake regulation and suggests that Hawaii's Act may draw upon or influence future state-level legislative efforts. While the Act does not explicitly amend or supersede specific existing statutes, it adds new sections to Chapter 11 of the Hawaii Revised Statutes, which pertains to elections, thereby integrating its provisions directly into the state's electoral legal framework.

International Alignment

As a state-level law in the United States, Hawaii Act 191 primarily focuses on domestic electoral integrity and does not directly aim for international alignment in the same way a national or supra-national regulation might. Its provisions are tailored to the specific legal and political context of Hawaii's elections. However, the underlying concerns that prompted its enactment—the spread of disinformation, the threat to democratic processes, and the malicious use of artificial intelligence—are global issues. Many countries and international bodies are grappling with similar challenges posed by deepfakes and generative AI. Therefore, while not explicitly designed for international alignment, Hawaii's Act contributes to the growing global discourse and regulatory landscape concerning AI governance and the protection of information integrity. The legislative findings explicitly mention that other U.S. states have enacted similar legislation, indicating an awareness of broader domestic legislative trends rather than international ones.

The principles embedded in Act 191, such as transparency through disclaimers and accountability for harmful content, resonate with broader international discussions on responsible AI development and deployment. International organizations, like the OECD, UNESCO, and the European Union, have been developing recommendations, guidelines, and regulations that emphasize transparency, explainability, and the mitigation of risks associated with AI, particularly in sensitive areas like elections and public discourse. While Hawaii's Act does not directly reference these international instruments, its focus on preventing voter manipulation and protecting electoral integrity through specific prohibitions and disclosure requirements aligns with the spirit of these global efforts to ensure AI is used ethically and responsibly. The challenges posed by deepfakes are inherently cross-border in the digital age, as content can originate anywhere and spread globally. Therefore, while the enforcement of Act 191 is confined to Hawaii's jurisdiction, its existence contributes to a wider body of legal precedents that could inform future international cooperation or the development of harmonized standards for deepfake regulation.

Implementation Timeline

Compliance Checklist

Sources and References