United States - Wisconsin - Deepfakes Regulation (Act 123)

Overview

Wisconsin Act 123, initially introduced as Assembly Bill 664, represents a significant legislative effort by the state of Wisconsin to address the burgeoning challenges posed by artificial intelligence (AI), particularly in the context of political communications and electoral integrity. The Act, which officially came into force on March 22, 2024, focuses specifically on regulating 'synthetic media'—commonly known as deepfakes—within political advertisements. This legislative initiative emerged amidst growing national and international concerns regarding the potential for AI-generated content to mislead voters, spread disinformation, and erode public trust in democratic processes. The rapid advancement of generative AI technologies has made it increasingly difficult for the average person to discern between authentic and fabricated audio or video content, necessitating clear regulatory frameworks to ensure transparency and accountability in political discourse.

The primary objective of Wisconsin Act 123 is to safeguard the integrity of elections by mandating explicit disclosures when AI-generated content is used in political advertising. Rather than outright banning deepfakes, the law adopts a transparency-first approach, requiring that voters be informed when they are encountering synthetic media. This approach is rooted in the belief that providing voters with accurate information about the origin of political content empowers them to make more informed decisions and critically evaluate the messages they receive. The Act targets communications paid for by various political entities, including candidate committees, legislative campaign committees, political action committees, independent expenditure committees, political parties, recall committees, and referendum committees, thereby covering a broad spectrum of electoral influence. The scope is carefully defined to address the most impactful forms of AI manipulation in political campaigns, acknowledging the need to balance free speech with the imperative for truthful and transparent communication.

Definitions

Central to Wisconsin Act 123 is the precise definition of 'synthetic media,' which the legislation establishes as 'audio or video content substantially produced by means of generative artificial intelligence.' This definition is crucial for delineating the scope of the law, clearly indicating that the regulation applies to content where AI plays a significant role in its creation or alteration. Generative artificial intelligence, in this context, refers to advanced AI techniques capable of autonomously creating new data, such as realistic audio or video, which can be difficult to distinguish from genuine recordings. The Act intentionally focuses on these sophisticated AI-driven manipulations, often termed 'deepfakes,' due to their high potential for deception and their capacity to convincingly portray individuals saying or doing things they never did.

It is important to note what the law does not regulate. The Act specifically differentiates between 'deepfakes' and 'manipulated media'—sometimes referred to as 'cheapfakes' or 'shallowfakes.' Manipulated media, which might involve simpler editing techniques like mislabeling videos or basic alterations to imply a different context, falls outside the purview of this legislation. Furthermore, the law does not extend to the regulation of deepfake photos. By narrowing its focus to AI-generated audio and video content in political communications, Wisconsin Act 123 aims to target the most technologically advanced and potentially misleading forms of synthetic media, while avoiding overreach into less sophisticated forms of media manipulation or other applications of AI-generated imagery. This targeted approach reflects a careful consideration of the specific threats posed by advanced AI in the electoral landscape.

Governance and Institutional Framework

The primary body tasked with the oversight and enforcement of Wisconsin Act 123 is the Wisconsin Ethics Commission. The Act explicitly grants the Commission rulemaking authority to implement its provisions, allowing for the development of detailed regulations and limited exceptions as needed. This delegation of authority is vital for ensuring that the law remains adaptable to the rapidly evolving landscape of AI technology and its applications in political campaigning. The Commission's role extends beyond mere enforcement; it is empowered to establish the practical guidelines and procedures necessary for compliance, which may include clarifying ambiguous terms or setting specific technical standards for disclosure. This proactive regulatory capacity is essential for maintaining the effectiveness of the Act in a dynamic technological environment.

The Wisconsin Ethics Commission is responsible for investigating alleged violations of the deepfake disclosure requirements and for administering the prescribed penalties. This institutional framework ensures a centralized and specialized approach to enforcing the integrity of political communications. The Commission's expertise in campaign finance and ethics provides a suitable foundation for addressing the novel challenges presented by AI-generated content. Furthermore, the Act specifies that any violation shall be enforced as provided under existing Wisconsin Statutes related to campaign finance enforcement, integrating this new regulation into established legal processes. This integration aims to streamline enforcement and leverage existing mechanisms for accountability in political activities, underscoring the state's commitment to maintaining fair and transparent elections.

Key Focus Areas

Wisconsin Act 123's core focus revolves around mandatory disclosure requirements for political communications that incorporate synthetic media. The law specifically targets audio and video content that is substantially produced by generative artificial intelligence and is used in communications paid for by designated political entities. These entities include candidate committees, legislative campaign committees, political action committees, independent expenditure committees, political parties, recall committees, and referendum committees. The intent is to capture a wide array of organized political messaging, ensuring that significant campaigns and advocacy efforts are subject to these transparency rules. The legislation aims to prevent the surreptitious use of AI to create misleading or fabricated narratives that could unduly influence voters without their awareness of the content's artificial nature.

The scope of regulated communications further extends to those containing 'express advocacy,' 'issue advocacy,' or those that support or oppose a 'referendum.' Express advocacy refers to clear and explicit calls for the election or defeat of a candidate, while issue advocacy pertains to communications providing information on political or social issues with the intent to influence election outcomes. By including these categories, the Act addresses both direct electoral campaigning and broader efforts to shape public opinion that could indirectly impact elections. This comprehensive approach ensures that various forms of politically influential content are covered, reinforcing the principle that transparency regarding AI-generated media is crucial across the spectrum of political discourse. The explicit inclusion of referenda also highlights the law's commitment to protecting the integrity of direct democracy processes.

Implementation Framework

The implementation framework for Wisconsin Act 123 is designed to ensure clear and unambiguous disclosure of synthetic media in political advertisements. For audio communications containing AI-generated content, the law mandates that the words 'Contains content generated by AI' must be included at both the beginning and the end of the communication. This dual placement ensures that listeners are alerted to the artificial nature of the content regardless of when they tune in or how much of the communication they hear. The requirement for specific wording aims to standardize disclosures, preventing vague or misleading labels that might undermine the transparency objective.

For video communications, the disclosure requirements are more nuanced to account for the visual and auditory elements. If a video includes only video synthetic media, the words 'This video content generated by AI' must be displayed. If it includes only audio synthetic media, the disclosure must state 'This audio content generated by AI.' In cases where both video and audio synthetic media are present, the comprehensive disclosure 'This content generated by AI' is required. Crucially, for video communications, these disclosures must be present 'throughout the duration of each portion of the communication containing synthetic media,' and must be 'readable, legible, and readily accessible.' This continuous and prominent display ensures that viewers are consistently aware of the AI-generated elements, preventing fleeting or easily missed disclaimers. The Wisconsin Ethics Commission is empowered to promulgate rules to further detail these implementation aspects, including potential limited exceptions, ensuring practical and effective application.

Monitoring and Evaluation

While Wisconsin Act 123 does not explicitly outline a separate, dedicated monitoring and evaluation framework, the ongoing role of the Wisconsin Ethics Commission inherently includes these functions. As the designated enforcement authority, the Commission is responsible for receiving complaints, investigating potential violations, and ensuring compliance with the disclosure requirements for synthetic media in political advertisements. This process of enforcement naturally involves monitoring the political communication landscape for instances where AI-generated content might be used without proper disclosure. The Commission's ability to promulgate rules also provides a mechanism for adaptive evaluation, allowing the regulatory framework to be updated or refined in response to new technological developments or observed challenges in implementation.

The effectiveness of the Act will largely depend on the vigilance of the public, political entities, and the media in reporting potential non-compliance, as well as the Commission's capacity to investigate and act upon such reports. The imposition of a clear penalty for violations serves as a deterrent and provides a measurable outcome for enforcement actions, which can be tracked over time to assess the law's impact. Furthermore, as AI technologies continue to advance, the rulemaking authority granted to the Ethics Commission becomes a critical tool for evaluating the adequacy of existing definitions and disclosure methods. This allows for necessary adjustments to ensure the law remains relevant and effective in achieving its goal of transparency in political communications, reflecting a dynamic approach to regulation in a rapidly evolving technological domain.

Penalties, Liability, and Appeals

Violations of Wisconsin Act 123's disclosure requirements for synthetic media in political advertisements are subject to specific penalties. The Act stipulates that whoever intentionally violates the provisions regarding audio or video disclosures shall be subject to a forfeiture not to exceed $1,000 for each violation. This civil penalty is designed to deter non-compliance and ensure accountability for political entities that fail to properly label AI-generated content. The enforcement of these forfeitures falls under the purview of the Wisconsin Ethics Commission, utilizing established procedures for enforcing campaign finance regulations. This integration into existing enforcement mechanisms aims to provide a clear and consistent process for addressing breaches of the Act.

Crucially, the Act also addresses the question of liability for various parties involved in the dissemination of political communications. It explicitly states that no liability for a violation of the disclosure requirements shall attach to any person who is a broadcaster or other host or carrier of a video or audio communication containing synthetic media, unless that person is also a committee responsible for the communication. This provision aims to protect media outlets and online platforms from being held responsible for content they merely transmit or host, placing the primary burden of compliance on the political committees that create and pay for the advertisements. Furthermore, the Act clarifies that its provisions may not be construed to alter or negate any rights, obligations, or immunities of a provider of an interactive computer service under 47 U.S.C. § 230, commonly known as Section 230 of the Communications Decency Act. This ensures that the state law does not inadvertently undermine federal protections for online platforms regarding third-party content.

Relationship to Other Instruments

Wisconsin Act 123 operates within the broader legal landscape of election law and technological regulation, establishing specific relationships with existing instruments. Primarily, it amends Wisconsin Statutes related to the attribution of political contributions, disbursements, and communications, integrating the new deepfake disclosure requirements into the established framework governing political advertising. This ensures consistency with existing regulations on transparency in political messaging and leverages established enforcement mechanisms under the Wisconsin Ethics Commission. The Act's focus on transparency through disclosure, rather than outright prohibition, aligns with constitutional considerations regarding free speech in political campaigns, aiming to inform voters without unduly restricting political expression.

A significant aspect of the Act's relationship to other instruments is its explicit reference to federal law, specifically 47 U.S.C. § 230. The Act states that its provisions may not be construed to alter or negate any rights, obligations, or immunities of a provider of an interactive computer service under this federal statute. Section 230 generally protects online platforms from liability for content posted by third parties, fostering an environment for free expression online. By including this clause, Wisconsin Act 123 acknowledges and respects the federal preemption in this area, ensuring that state-level deepfake regulations do not inadvertently create new liabilities for online platforms that host political advertisements. This careful drafting reflects an awareness of the complex interplay between state and federal law in the digital sphere, particularly concerning content moderation and platform responsibility.

International Alignment

While Wisconsin Act 123 is a state-level initiative within the United States, its enactment reflects a broader global trend towards regulating artificial intelligence, particularly deepfakes, in sensitive areas like elections. Numerous countries and sub-national jurisdictions worldwide are grappling with similar challenges posed by generative AI and its potential to disrupt democratic processes. The Act's approach of mandating transparency through disclosure aligns with a common regulatory strategy seen in other jurisdictions, which often prioritize informing the public over outright bans, especially in contexts protected by free speech principles. This legislative development positions Wisconsin as one of several U.S. states, alongside others like California, Michigan, Minnesota, and Texas, that have moved to address AI-generated content in political campaigns, indicating a growing consensus on the necessity for such safeguards.

The specific focus on political advertisements and the requirement for clear labeling resonate with discussions at international forums and by global bodies concerning responsible AI governance. Although there isn't direct "international alignment" in the sense of a treaty or mutual recognition agreement, Wisconsin's law contributes to the global dialogue on how to manage the societal impacts of AI. It demonstrates a shared recognition of the risks that synthetic media poses to public trust and democratic integrity, fostering a regulatory environment where truthfulness and source attribution are increasingly valued. The challenges of cross-border dissemination of deepfakes, however, highlight the limitations of purely national or sub-national regulations, suggesting a continued need for international cooperation and harmonized approaches to fully address the global nature of AI-driven disinformation.

Implementation Timeline

Compliance Checklist

Sources and References