By: William H. Kronblat

A family member sends you a video featuring a prominent political figure saying some rather crazy things. Your family member is in disbelief over what the political figure said. You are also shocked at the contents of the video, but something seems off. You are not sure, but you think the video has been edited or altered with artificial intelligence. You check the video’s comments and it turns out other people were also shocked by the video’s contents, but others also suspected the video may have been edited. Turns out the video was not a real speech at all, but a deepfake.

What are Deepfakes? What’s the Concern?

As social media usage increases and as social media becomes a predominant news source for many individuals, a new problem arises. That problem is deepfakes. Deepfakes are  “online imagery that can make anybody appear to do or say anything within the limits of one’s imagination, cruelty, or cunning.” Deepfake technology has continued to advance, and the proliferation of deepfakes is growing at an alarming rate. The World Economic Forum reported that in 2020, the number of online deepfake videos was approximately  145,277; a sizable jump from the 14,678 deepfake videos reported the previous year. 

United States agencies are feeling uneasy over the  rapid increase in online deepfakes. For instance, the NSA, FBI, and CISA, have expressed concerns on the proliferation of deepfake technology, noting that malicious actors may use deepfake technology to perpetuate “disinformation operations.” Similarly, the Department of Homeland Security identified deepfakes as a potential national security concern and as a threat to “individuals and industries.”

State Action on Deepfakes:

Given the growing concerns surrounding deepfake technology, you may be asking, what is being done to address this issue? What are government agencies or elected officials doing to stop deepfake misinformation and election disinformation? 

California has been a leader amongst U.S. states in terms of regulating AI and deceptive AI campaign ads. California has passed various pieces of legislation aimed at deterring deceptive campaign ads including a law that requires political action committees to disclose the use of artificial intelligence to create or alter ads. California Governor, Gavin Newsome, has mentioned deepfakes relating to election disinformation as a particular area of state concern. 

Ahead of the 2024 election, Gavin Newsom signed three bills aimed at addressing AI generated false images or videos in political ads. One of the bills, Assembly Bill 2839, particularly focused on deepfakes. Assembly Bill 2839 would have “made it illegal to distribute ‘materially deceptive audio or visual media of a candidate’ 120 days before an election and, in some cases, 60 days after.” 

Current State of Assembly Bill 2839:

Assembly Bill 2839 was set to take effect immediately, but it has run into a legal hurdle.   Federal U.S District Judge John A. Mendez granted a preliminary injunction enjoined Assembly Bill 2839 following a lawsuit to block the legislation. In his decision, Judge Mendez found that Assembly Bill 2839 likely violates the First Amendment and is unconstitutional because it did not pass the strict scrutiny standard for regulating speech content. Judge Mendez gave substantial weight to California’s interest in protecting fair elections, but noted that “this interest must be served by narrowly tailored ends.”  Judge Mendez cited to the US Supreme Court’s decision in New York Times v. Sullivan, where the Supreme Court affirmed that even “lies with malice” were protected speech under the First Amendment. Judge Mendez further clarified that the concerns surrounding deepfakes “do not give legislators unbridled licenses to bulldoze over the longstanding tradition of critique, parody, and satire protected by the First Amendment.”

Christopher Kohls, a conservative commentator, brought the lawsuit to block the legislation. Kohls is responsible for numerous deepfake videos satirizing Democrats, including Presidential nominee Vice President Kamala Harris. Governor Gavin Newsome cited one of Kohls’ videos when he signed Assembly Bill 2839, but the Bill was introduced in the California Legislature well before Kohls published the video Newsome referred to. Kohls drew support from Elon Musk, the owner of X, who boosted Kohls’ video and criticized Assembly Bill 2839 claiming that Gavin Newsom was trying to “make parody illegal.” 

What’s Next?

It is important to note that Judge Mendez only issued a preliminary injunction, which is merely temporary. In other words, this injunction does not permanently get rid of Assembly Bill 2839. The preliminary injunction simply blocks the law from going into effect until Christopher Kohls’  lawsuit is completed.  However, Judge Mendez said that Kohls is likely to succeed in his lawsuit and get Assembly Bill 2839 struck down as unconstitutional because the Bill is too broad and runs afoul of First Amendment free speech protections. Judge Mendez suggested that Assembly Bill 2839 was excessive and that a less restrictive approach like “counter speech” would be more permissive. However, Judge Mendez did not address this concept in detail.

While California’s legislation on deepfakes is facing an uphill battle, federal legislation aimed at addressing election deepfakes has been proposed in Congress by a bipartisan of lawmakers. This Bill would allow the Federal Trade Commission to regulate AI generated election content similarly to how it has regulated other political misrepresentations for decades. 

The Federal Election Commission has also started to consider potential regulations aimed at curbing AI and deepfake election-related misrepresentations. The FEC has already taken some steps at addressing AI generated content when they outlawed AI-generated robocalls that were meant to discourage voters. 

These proposed regulations still have to be approved and also clear the First Amendment hurdles that Assembly Bill 2839 has run into. Nonetheless, it is clear that Assembly Bill 2839 will not be the only attempt to regulate deepfakes and misinformation, at least in the election setting.  The concerns surrounding deepfakes do not appear to be trivial, but finding a solution that respects the fundamental protection of the first protection is easier said than done. 

 #Deepfakes, #First Amendment, #Misinformation