Ahead of the US Presidential Election, AI-Generated Fake Photos Surge... State Governments Begin Deepfake Regulations in Earnest

Measures to Prevent Damage Prepared by State Legislature
Federal Law Application Unclear... Biden "Risk Management Needed"

Last month, a photo showing an explosion incident inside the Pentagon, the U.S. Department of Defense headquarters, circulated, causing a temporary drop in the U.S. stock market. Earlier this month, a photo of former U.S. President Donald Trump having physical contact with Dr. Anthony Fauci, former director of the National Institute of Allergy and Infectious Diseases, with whom he had conflicts during the COVID-19 crisis, drew attention. Both photos were fake images created by artificial intelligence (AI), so-called 'deepfake' photos, making it appear as if events that never happened actually occurred.

Bloomberg News reported on the 20th (local time) that U.S. states are preparing measures to regulate deepfake photos. With the presidential election next year approaching, the misuse of generative AI to create deepfake photos has become a growing social problem, increasing the need for regulation. Consequently, state legislatures have begun the process of drafting related bills.

This is a photo of the Pentagon, the United States Department of Defense, taken in March of last year. [Image source=AP Yonhap News]

A photo capturing former U.S. President Donald Trump (left) and former National Institute of Allergy and Infectious Diseases Director Anthony Fauci passing each other was actually taken in March 2020.
[Image source=Reuters Yonhap News]

According to the report, nine U.S. states currently have laws regulating deepfakes related to pornography and elections. California, Texas, and Virginia enacted deepfake-related laws as early as 2019, before the AI boom. Minnesota passed related legislation in May, and Illinois has passed a similar bill through its legislature, with only the governor's final signature remaining. Four other states are currently undergoing legislative procedures on this matter.

Deepfake refers to technology that synthesizes human images using AI. The term originated in 2017 on the U.S. online community Reddit, where a user uploaded fake pornographic videos by applying AI algorithms to place celebrities' faces without consent. The number of deepfake photos and videos has rapidly increased alongside AI technology development.

Deeptrace Labs, a service that detects deepfakes, pointed out that about 15,000 deepfake videos were identifiable online in 2019, with 96% featuring women in adult content. Another deepfake detection company, Sensity AI, also noted an exponential increase in deepfake videos since 2018.

Especially this year, interest in generative AI has grown, and cases of harm have increased, but discussions in the federal Congress remain sluggish. Bloomberg reported that state legislatures are accelerating the drafting of bills to immediately prevent damage. The Hill, a U.S. political media outlet, recently analyzed that with the upcoming U.S. presidential election next year, deepfake photos are expected to significantly change election campaign methods.

[Image source=Reuters Yonhap News]

Experts point out that it is difficult to file lawsuits against deepfake photos under current laws. Rebecca Delfino, a law professor at Loyola Marymount University who researches deepfakes, said, "As technology continues to develop, it is very difficult for anyone other than digital forensic experts to distinguish what is fake and what is real." Matthew Kugler, a professor at Northwestern University, noted, "If you take something fully public like a face, there is no certain way for people to sue under current laws."

Currently, it may be difficult to file lawsuits without specific laws related to pornographic deepfakes. Although 48 states in the U.S. prohibit revenge porn and hidden camera footage, Professor Kugler expressed concern that these laws might not apply considering deepfakes are fake. There are also concerns that defamation due to portrait rights infringement may only apply commercially in some states, making it distant from deepfake-related punishments.

The varying scope of laws enacted by each state and differing definitions of deepfakes are identified as problems. Experts also worry that since deepfakes can be used in ways that do not harm others, it is difficult to prohibit harmful deepfakes while protecting legal deepfakes.

In fact, the Wall Street Journal (WSJ) recently reported that celebrities are beginning to directly contract regarding their images or videos synthesized or altered using AI. WSJ also noted that AI technology can make celebrities appear younger by lowering their apparent age or create photos or videos showing them doing things they actually cannot do.

Meanwhile, U.S. President Joe Biden met with civic leaders and experts in San Francisco, California, on the same day and pledged to manage AI risks. He said, "We will witness more technological changes in the next 10 years than in the past 50 years, or perhaps even longer." He emphasized, "We need to manage risks to society, the economy, and national security. From privacy protection to addressing AI bias and fake news, and ensuring AI systems are safe before release, we are committed to protecting the rights and safety of Americans."

This content was produced with the assistance of AI translation services.