Hank Stephenson’s British Standard detector is appropriately regulated. The seasoned journalist has built a career out of finding political propaganda and misinformation. However, when he first viewed a clip of one of his state’s most well-known candidates for Congress, even he was initially tricked. On his phone screen, Kari Lake, the Republican nominee for Senate from Arizona, was speaking in words prepared by a program developer. Stephenson was viewing a deepfake, a film created by his media business, Arizona Agenda, using artificial intelligence to highlight the risks associated with disinformation from this technology during a crucial election year.
The website’s co-founder, Stephenson, stated in an interview that he was “blown away” by what he saw when they started doing this. “I assumed it would turn out to be so horrible it might not mislead anyone,” Stephenson said. We also need more sophistication. It is problematic because if we can pull it off, anyone with a genuine budget can do it well enough to fool you and me.
Experts and policymakers are increasingly warning about the possible disastrous impact of artificial intelligence deepfakes as the close of the 2024 presidential election approaches. They believe this technology might further damage the nation’s perception of truth and disrupt the electorate.
There are indications that artificial intelligence and the anxiety surrounding it are already impacting the race. The makers of an advertisement featuring the well-known public blunder mistake of the previous president, Donald Trump, were wrongly accused of trading in artificial intelligence (AI) stuff late last year. Meanwhile, real-life Photoshop-manipulated photos of Trump and other politicians, intended to both charm and humiliate, have repeatedly gone viral, causing chaos during a pivotal moment in the election campaign.
Officials are now hurrying to react. Recently, the New Hampshire Department of Justice declared that it was looking into a fake robocall that used President Biden’s artificial intelligence (AI) voice; the state of Washington alerted voters to the dangers of deepfakes; and legislators from Oregon to Florida implemented laws limiting the application of this technology in political communication.
Additionally, the senior elections official in Arizona—a crucial battleground state in the 2024 election—used deepfakes of himself as part of an instructional session to prepare staff members for the flood of lies that lay ahead. The task motivated Stephenson and his associates at Arizona Goals, a daily publication dedicated to clarifying intricate political tales for a readership of roughly ten thousand subscribers.
They enlisted the assistance of a tech-savvy buddy and explored ideas for around a week. Stephenson released the article on Friday and three deepfake videos of Lake.
The article opens with a deceptive statement informing readers that Lake, a hard-right politician whom the Arizona Agenda has previously criticized, has endorsed the publication and praises how much she loves it. However, the film soon turns into an obvious joke.
Get the Arizona Agenda by subscribing to get brutal real news—not the phony “And an indication of the terrible artificial intelligence heading to you in the coming elections, such as this clip, which is an artificial intelligence fake information the Arizona Agenda developed to show you just how amazing this technology is going,” Lake says to the camera before continuing.
Tens of thousands of people had watched the films by Saturday. The actual Lake, whose campaign lawyers had issued the Arizona Agenda a cease-and-desist letter, was not pleased. The previously mentioned deepfake videos were ordered to be taken down immediately from all platforms they had been shared or distributed. The letter threatened to utilize all legal means at Lake’s campaign’s disposal if the media did not cooperate.
When approached on Saturday, a campaign representative was unwilling to respond.
Stephenson stated late Saturday afternoon that he had not planned to take down the recordings but was still discussing his response options with legal counsel. He claimed that the deepfakes are helpful educational tools and that he wants to provide readers with the ability to recognize these fakes before the elections heat up and they are overwhelmed with them.
In the article accompanying the video clips, Stephenson stated that it is up to all of us to fight the latest electronic misinformation during the current election season. Understanding what is available to you and applying critical thinking are your best lines of defense.
According to Hany Farid, an educator at the University of California, Berkeley, who specializes in studying electronic propaganda and disinformation, the Arizona Agenda videos were helpful PSAs that were thoughtfully produced to minimize unexpected penalties. Nonetheless, he warned media outlets to exercise caution when presenting their deepfake reporting.
Farid stated, “I appreciate the PSAs; however, there is moderation.” You don’t want viewers and readers to assume that everything conflicting with their expectations is phony.
According to Farid, deepfakes offer two different “risk carriers.” First, dishonest actors can fabricate recordings of individuals expressing stuff that was never stated, and second, people are more likely to believe that any genuine embarrassing or implicating video was created.
According to Farid, this dynamic has been particularly noticeable during the conflict that is filled with misinformation—Russia’s attack on Ukraine. In the early stages of the conflict, Ukraine circulated a deepfake image of Paris being attacked, calling on international leaders to respond to the actions of the Kremlin with the same enthusiasm as they would have if the Tower of Paris were attacked.
Farid stated that although it was a strong statement, it gave rise to Russia’s unjustified claims that later records from Ukraine, which revealed proof of war crimes committed by the Kremlin, were also produced.
He said, “I’m concerned that all of it is turning suspicious.”
Stephenson shares a similar fear; his neighborhood is a political battlefield that has recently become a center of false allegations and conspiratorial ideas.
He remarked that we have been fighting over what is genuine for a long time. While objective videos will now be brushed off for being fake and treated as reality, truthful information can now be dismissed as fake news.
Researchers such as Farid are working nonstop to develop algorithms to make it easier for journalists and other users to identify deepfakes. Farid added that the Arizona Agenda film was quickly identified as a fake by the technologies he presently employs. It is encouraging, given the wave of fakes that will likely come. However, deepfake technology is developing quickly, so future fakes might be more challenging to identify.
Although Stephenson’s poor deepfake tricked a couple of individuals, a few paying subscribers unsubscribed when we sent out Friday’s newsletter titled “Kari Lake does us a solid.” Stephenson speculates that they most likely believed Lake’s recommendation to be genuine.
