Deepfake laws in the United States comprise a patchwork of federal and state statutes enacted to regulate the creation, distribution, and use of synthetic media generated by artificial intelligence systems. These laws address a range of harms associated with deepfakes, including election interference, nonconsensual intimate imagery, fraud, and defamation. The legal landscape has evolved rapidly since the mid-2010s, driven by advances in generative AI and the proliferation of accessible deepfake creation tools. As of 2025, the majority of U.S. states have enacted some form of deepfake-related legislation, though the scope and enforcement mechanisms vary considerably.
Deepfakes are a portmanteau of 'deep learning' and 'fake', referring to images, videos, or audio edited or generated using machine learning techniques. The technology leverages neural networks, including variational autoencoders and generative adversarial networks (GANs), to create highly realistic synthetic media. While the act of creating fake content predates digital technology, deepfakes uniquely employ AI-based tools to produce media that can depict real or fictional people with unprecedented fidelity. This capability has raised significant legal and policy questions about privacy, intellectual property, and democratic integrity.
Federal Legislative Efforts
At the federal level, Congress has considered multiple bills addressing deepfakes but has not enacted comprehensive legislation as of 2025. The Deepfake Accountability Act, first introduced in 2019, proposed requiring digital watermarks on AI-generated content and criminal penalties for failing to disclose synthetic media. The Identifying Outputs of Generative Adversarial Networks Act, signed into law in 2020, directed the National Science Foundation and the National Institute of Standards and Technology to support research on deepfake detection and content authenticity.
The DEEPFAKES Accountability Act, reintroduced in several sessions, would have required creators of deepfakes to embed digital watermarks and disclose the synthetic nature of their content. The bill also proposed creating a private right of action for individuals depicted in deepfakes without consent. However, these federal proposals have faced challenges in reconciling free speech protections with regulatory objectives, and none have advanced to a final vote in both chambers.
In 2023 and 2024, federal attention shifted toward election-related deepfakes. The proposed Protect Elections from Deceptive AI Act would have prohibited the distribution of materially deceptive AI-generated audio or visual media depicting federal candidates within 120 days of an election. This bill reflected growing concern about the potential for deepfakes to interfere with electoral processes, a risk that academics and policymakers have highlighted alongside concerns about disinformation and hate speech.
State-Level Regulation: Overview
State legislatures have been more active than Congress in addressing deepfakes. By 2025, over 40 states had enacted laws related to synthetic media, with most focusing on specific harms rather than comprehensive regulation. The first wave of state laws, beginning around 2019, targeted nonconsensual intimate imagery (NCII), often referred to as revenge porn, which had already been criminalized in many states before deepfake technology became widespread.
California emerged as a pioneer in deepfake regulation. In 2019, the state enacted Assembly Bill 730, which prohibited the distribution of materially deceptive audio or visual media of a candidate for elected office within 60 days of an election. The same year, California passed Assembly Bill 602, which criminalized the creation and distribution of deepfake pornography without consent. These laws established a template that other states would follow, combining election integrity provisions with protections against intimate image abuse.
Nonconsensual Intimate Imagery Laws
The most widely adopted category of deepfake legislation addresses nonconsensual intimate imagery. As of 2025, nearly every state has laws that explicitly include digitally altered or AI-generated intimate images within their definitions of NCII. These statutes typically impose criminal penalties for creating or distributing deepfake pornography that depicts a real person without their consent, with penalties ranging from misdemeanors to felonies depending on the jurisdiction and circumstances.
Texas enacted one of the more stringent laws in 2019, making it a felony to create or distribute deepfake intimate images with intent to harass or harm. The law also provides civil remedies, allowing victims to seek damages and injunctive relief. New York followed in 2022 with a law that criminalized the dissemination of synthetic intimate images, while also establishing a civil cause of action. These laws reflect recognition that deepfake NCII can cause severe psychological, reputational, and economic harm to victims.
Some states have extended NCII laws to cover threats to create deepfake intimate images. For example, Illinois amended its existing revenge porn statute to include threats, recognizing that the mere threat of creating a deepfake can be used as a tool of coercion and control. This expansion addresses the unique nature of deepfakes, which can be created without any original intimate imagery of the victim.
Election and Political Deepfake Laws
Election-related deepfake laws constitute the second major category of state regulation. These laws typically prohibit the distribution of deceptive synthetic media depicting political candidates within a specified period before an election, often 60 or 90 days. Most such laws require that the content be materially deceptive, meaning it would cause a reasonable person to believe the candidate said or did something they did not actually say or do.
California's AB 730, enacted in 2019, was the first law of its kind. It allowed candidates depicted in deceptive deepfakes to seek injunctive relief and imposed civil penalties. Minnesota passed a similar law in 2023, and Michigan, Washington, and Wisconsin followed with their own versions. By 2024, at least 20 states had enacted election-related deepfake laws, with many taking effect before the 2024 presidential election.
These laws have faced First Amendment challenges. In 2024, a federal court in Florida struck down portions of that state's deepfake election law, finding that it was overbroad and chilled protected speech. The ruling highlighted the tension between preventing electoral deception and preserving political speech, a tension that continues to shape legislative drafting in other states. Some states have responded by narrowing their definitions of deceptive content or adding exemptions for satire, parody, and news reporting.
Civil Remedies and Right of Publicity
Beyond criminal statutes, many states have expanded civil remedies for deepfake victims through right of publicity laws. These laws protect individuals' control over the commercial use of their name, image, and likeness. Tennessee, in 2024, became the first state to explicitly include AI-generated replicas of performers' voices and images within its right of publicity statute, known as the Ensuring Likeness Voice and Image Security (ELVIS) Act. The law was prompted by concerns from the music industry about AI-generated songs that imitated artists' voices without authorization.
California and New York have also amended their right of publicity laws to address digital replicas. California's 2020 amendment to its Civil Code clarified that liability extends to "digital replicas" created through AI. New York's 2020 law, which took effect in 2021, created a post-mortem right of publicity that includes digital replicas of deceased performers. These laws provide a civil cause of action for damages, allowing victims to sue for unauthorized commercial use of their likeness in deepfakes.
Fraud and Financial Crime Provisions
Some states have addressed deepfakes in the context of fraud and financial crimes. These laws recognize that deepfakes can be used to impersonate individuals for financial gain, such as through voice cloning to authorize fraudulent transactions or video deepfakes to deceive employees into transferring funds. As of 2025, at least a dozen states have enacted laws that specifically address AI-enabled fraud, either by expanding existing fraud statutes or creating new offenses.
Arizona and Louisiana have enacted laws that enhance penalties for crimes committed using deepfake technology. These statutes treat the use of synthetic media as an aggravating factor, increasing the severity of charges for underlying offenses such as identity theft, wire fraud, or extortion. This approach allows prosecutors to apply existing criminal frameworks while recognizing the unique threat posed by AI-generated deception.
Detection and Disclosure Requirements
A growing number of states have enacted laws requiring disclosure of AI-generated content. These laws typically mandate that deepfakes be labeled or watermarked to inform viewers of their synthetic nature. California's 2024 law, effective in 2025, requires large online platforms to label AI-generated content that could deceive viewers about elections or public health. The law also requires platforms to provide tools for users to report unlabeled deepfakes.
Utah and Colorado have enacted similar disclosure laws, with Colorado's 2024 statute requiring political advertisements to disclose the use of generative AI. These disclosure requirements aim to address the informational harms of deepfakes by empowering viewers to make informed judgments about the authenticity of media. However, enforcement mechanisms vary, and some laws rely on platform self-regulation rather than government oversight.
Enforcement Challenges and Criticisms
Enforcement of deepfake laws faces significant challenges. The decentralized nature of the internet makes it difficult to identify and prosecute creators who may be located outside the jurisdiction where the harm occurred. Additionally, the rapid evolution of large language models and other AI technologies means that detection tools often lag behind generation capabilities. Image forensics researchers have developed techniques to detect manipulated images, but these methods are not always reliable against the most sophisticated deepfakes.
Critics of deepfake laws raise concerns about overbreadth and chilling effects on legitimate speech. Artists, journalists, and satirists have argued that some statutes could be used to suppress protected expression, such as political parody or documentary films that use synthetic media. The American Civil Liberties Union and other advocacy groups have filed amicus briefs in several cases challenging deepfake laws on First Amendment grounds, arguing that the government's interest in preventing deception does not justify broad restrictions on content.
Another criticism concerns the inconsistency of state laws, which creates a complex compliance environment for platforms and creators. A deepfake that is legal in one state may be criminal in another, and the lack of federal uniformity complicates enforcement across state lines. Some legal scholars have called for a federal framework that establishes minimum standards while allowing states to enact stricter protections.
Future Directions
As of 2025, the legal landscape for deepfakes in the United States remains in flux. Federal legislation has been proposed but not enacted, and state laws continue to evolve in response to technological developments and court challenges. The Supreme Court has not yet ruled on the constitutionality of deepfake laws, leaving lower courts to grapple with the balance between regulation and free expression.
Emerging issues include the use of deepfakes in the entertainment industry, where OpenAI and other companies have developed tools that can generate realistic video and audio. The potential for deepfakes to disrupt media industries has prompted calls for industry self-regulation and technical standards for content provenance. The Coalition for Content Provenance and Authenticity, an industry group, has developed technical specifications for cryptographically signed metadata that can verify the origin and history of digital content.
Legislators in several states have introduced bills for the 2025-2026 session that would expand protections for individuals depicted in deepfakes, including provisions for expedited takedown procedures and enhanced damages for malicious use. Some proposals would require platforms to implement deepfake detection tools and remove flagged content within a specified timeframe. These proposals reflect ongoing efforts to adapt legal frameworks to the challenges posed by AI-generated synthetic media.