Welcome to the Thursday, June 27, Brew.
By: Mercedes Yanora
Here’s what’s in store for you as you start your day:
- Ballotpedia launches Artificial Intelligence Deepfake Legislation Tracker
- Washington state Sen. Joe Nguyen and Oklahoma state Rep. Toni Hasenbeck discuss their deepfake legislation on a special two-part episode of On the Ballot
Ballotpedia launches Artificial Intelligence Deepfake Legislation Tracker
Deepfakes, an offshoot of artificial intelligence (AI), have become a pressing social and political issue that an increasing number of state lawmakers are trying to address through legislation. The number of bills has increased 950% in 2024. From 2019 to 2023, an average of 28 bills were introduced per year in state legislatures. In 2024, that figure jumped to 294.
That’s why we’ve created and today are officially launching a comprehensive AI Deepfake Legislation Tracker. As part of that launch, we have published Ballotpedia’s State of Deepfake Legislation 2024 Annual Report, available here.
The goal of the tracker is simple: to let people know what’s happening—in real time—with deepfake legislation in all 50 states. The tracker provides historical context on deepfake legislation going back to 2019, and covers topics such as:
- Political and election-related communications;
- Pornographic and child sexual abuse material;
- Fraud and extortion; and
- Property rights.
Not familiar with using legislation trackers? No problem. Click here to learn how to use the tracker.
Now, let’s take a closer look at both the deepfake policy landscape and current legislation.
Defining deepfakes
Deepfakes are videos, images, or audio files that artificial intelligence has generated or manipulated to realistically portray something that did not actually occur. According to the U.S. Government Accountability Office, “Deepfakes rely on artificial neural networks, which are computer systems modeled loosely on the human brain that recognize patterns in data. Developing a deepfake photo or video typically involves feeding hundreds or thousands of images into the artificial neural network, ‘training’ it to identify and reconstruct patterns—usually faces.”
Concerns surrounding deepfakes
Public policy organizations, lawmakers, and journalists have expressed concerns about deepfake technology. These concerns include the proliferation of child sexual abuse material and other nonconsensual sexual content, the distribution of deceptive political communications and election disinformation, property rights infringement, harassment, fraud, and threats to national security.
Policy approaches vary as lawmakers and advocates balance concerns about deepfakes with First Amendment free speech considerations and beneficial uses of the technology. The Foundation for Individual Rights and Expression writes, “Any government restriction on the expressive use of AI needs to be narrowly tailored to serve a compelling governmental purpose, and the regulation must restrict as little expression as is necessary to achieve that purpose.”
Current deepfake policy across the U.S.
As of June 21, 29 states had enacted laws related to pornographic deepfakes. Some of those laws specifically address the creation and distribution of child sexual abuse material, while others address the nonconsensual creation and distribution of adult intimate images. Twelve states have passed laws addressing both.
As of June 21, 17 states had enacted laws related to political deepfakes. In 10 of those states, the law applies to political materials distributed within a certain number of days before an election.
In 15 of the 17 states that have passed laws regarding political deepfakes, the law makes an exception for materials containing a disclosure statement, with varying degrees of specificity regarding what that statement must say and how it must be presented. Laws in Minnesota and Texas do not make disclosure exceptions.
Legislative activity in 2024
The chart below shows the number of deepfake-related bills enacted between 2019 and 2024, with 60% enacted in 2024 alone.
The following chart shows the number of deepfake-related bills introduced between 2019 and 2024. Sixty-seven percent were introduced in 2024.
Forty-four states considered bills related to deepfakes in 2024, with the most—36—in New York.
Four states—Montana, Nevada, North Dakota, and Texas—have legislatures that only meet in odd-numbered years.
Democrats sponsored the majority of deepfake-related bills (127), followed by bipartisan sponsors (111) and Republican sponsors (80).
Legislators introduced 46% of deepfake-related bills in states with a Democratic trifecta, 38% in states with a Republican trifecta, and 16% in states with divided governments.
The following chart shows the four most common topics of deepfake-related bills.
The following chart breaks those topics down by the party affiliation of the bills’ sponsor(s):
Washington state Sen. Joe Nguyen and Oklahoma state Rep. Toni Hasenbeck discuss their deepfake legislation on a special two-part episode of On the Ballot
Today, we are releasing two episodes of On the Ballot, Ballotpedia’s weekly podcast. Both episodes complement the launch of our new AI Deepfake Legislation Tracker.
In these episodes, Ballotpedia Editor-in-Chief Geoff Pallay interviews two legislators whose bills regulating deepfakes were approved this year.
- In the first episode, Washington state Sen. Joe Nguyen (D-34) discusses his sponsorship of WA SB 5838, which established an artificial intelligence task force that will explore different use cases, think through how to address developments like deepfakes, and set Washington’s policy course on AI. The conversation also touches on what motivated Nguyen and his colleagues to propose a task force, the bill-writing process, bipartisan cooperation, and policy predictions at the state and national level.
- In the second episode, Pallay interviews Oklahoma state Rep. Toni Hasenbeck (R-65) about her sponsorship of several AI bills. They discuss OK HB 3642 at length. The bill targets deepfakes by updating the state’s laws on obscenity and child pornography to include AI-generated images. They discuss what originally made her interested in this type of legislation, the bill-writing process, how to legislate something that is continuously evolving, and policy predictions at the state and national level.
If you would like to learn more about deepfake legislation, explore our tracker here and subscribe to On the Ballot on YouTube or your preferred podcast app! New episodes of On the Ballot drop every Thursday.
Listen here