By Giavanna Ventura.
The rise of deepfake technology has sparked debate among lawyers, scholars, and politicians due to the challenge of reconciling regulation with the First Amendment. Deepfakes use artificial intelligence software to create realistic videos, audio recordings, and images. While they were initially used sparingly for entertainment, the quality and accessibility of deepfakes have steadily evolved, leading to concerns over misinformation and privacy violations. One issue with regulating deepfake technology is that any laws regulating the use of deepfakes must also respect the First Amendment, and Arizona is one jurisdiction currently struggling with this issue.
The State of Deepfake Regulations in Arizona
To combat the dangerous potential of deepfakes, the Arizona House and Senate proposed legislation to penalize the use of deepfakes and provide remedies for people who are the subject of deepfakes posted without their consent. The legislature presented four major bills: HB2394, SB1359, SB1078, and SB1336. Senate Bill 1336 faced pushback from House Democrats, at least due in part to First Amendment concerns. It was vetoed by Gov. Hobbs, as she deemed it duplicative of Senate Bill 1078, but without Senate Bill 1078’s language focusing on the intent to defraud or harass.
Senate Bill 1078 walks the line between deepfake protection and freedom of speech. The bill amends Arizona Revised Statutes Section 13-2006, which focuses on criminal impersonation. It added felony penalties for the use of computer generated images, voice recordings, or videos of a depicted person with intent to defraud or harass. There are exceptions for parody, artistic expression, and any situations where it is clear to a reasonable audience that the content has been digitally altered.
Republican Senator John Kavanagh sponsored Senate Bill 1078 in Arizona 56th Legislature Second Regular Session, and the session ended June 15, 2024. Since SB1078 did not pass concurrence in the Senate before the end of the session, it is dead. The bill could be re-proposed due to its popularity with Gov. Hobbs, but its future remains unclear.
Presently, two of the four bills relating to deepfakes have passed: House Bill 2394 and Senate Bill 1359. House Bill 2394 allows a person running for office or any other citizen of Arizona to bring a legal action for digital impersonations posted without the depicted person’s consent and containing no indication that it is a digital impersonation to a reasonable person. This notably includes modified and AI-generated media such as deepfakes. In certain cases, the plaintiff can receive damages and injunctive relief. The bill also sets out situations where the aggrieved party can receive a preliminary declaratory injunction—namely, if the person is a political candidate or candidate for public office.
Senate Bill 1359 also targets deepfakes of candidates for elected office. It added Section 16-1023 to the Arizona Revised Statutes. Within 90 days of an election, any person who creates and distributes a deepfake depicting a candidate on the ballot must include a clear and conspicuous disclosure that the media contains content created by artificial intelligence. Failure to comply will result in civil penalties. Like the other legislation, this bill also includes exceptions for comedy, parody, and artistic expression.
What This Means
So far, no Arizona laws criminalize the use of deepfakes. Even proposed laws that target deepfakes allow for exceptions for satire, parody, and artistic expression. Arizona’s laws only target malicious uses of deepfakes and provide outs to avoid penalties, such as AI labeling to avoid penalties. There is a growing belief that rather than relying solely on regulation of deepfakes, educating the public further and raising media literacy could be a more effective strategy. Similarly, criminalizing the use of deepfakes may halt further innovation in the field without providing any substantial benefit.
Arizona’s current regulations, which provide injunctive and monetary relief, are broad enough to regulate, yet likely narrow enough to avoid First Amendment violations. Additionally, House Bill 2394 and Senate Bill 1359 offer extra protection for candidates for elected office. The main goal of these bills is to stop the spread of misinformation for voters, while also maintaining the right to free speech and expression.
Arizona’s work battling deepfakes does not end with regulation. Arizona Secretary of State Adrian Fontes spearheaded a training exercise for election workers about the threat of AI and deepfakes, including an AI-generated video of himself as evidence of their persuasiveness. Other education efforts include educating reporters and law enforcement officials.
Although legislators crafted the passed legislation to avoid infringing on the First Amendment, it remains uncertain whether the bills pass constitutional muster. This will likely remain unclear until there are significant challenges or case law on the matter.
Gia Ventura is a 2L at the Sandra Day O’Connor College of Law and is focusing her studies on patent and IP law. Over the summer of 2024, Gia worked as a legal intern at Skysong Innovations. She grew up in Austin, Texas, where she graduated high school before she and her family relocated to Phoenix, and earned a B.S. in Chemistry from ASU. During COVID, she studied to become a pharmacy technician before deciding to pursue a career in law. She looks forward to leveraging her technological background to provide valuable perspectives on intellectual property law through ASLJ.