With the U.S. Presidential race heating up, it’s not just U.S. adversaries who are trying to influence the campaign – even the candidates and their supporters are engaged in disinformation.
On a day when the top U.S. intelligence and cybersecurity agencies came together to identify Iran’s role in cyberattacks and influence operations aimed at gaining sensitive information on the U.S. elections and trying to tip the scales toward their favored presidential candidate, Donald Trump himself resorted to spreading deepfakes from AI-generated images and videos of Taylor Swift and Elon Musk.
Trump’s Deepfakes: ‘I Accept!’
U.S. Republican presidential candidate Donald Trump has raised a storm on social media after posting what looks like an AI-generated deepfake of Taylor Swift and her “swifties” calling for support for Trump’s election campaign.
On Sunday, Trump posted a collage of several tweets taken from X, formerly known as Twitter, with a caption “I accept!” on his Truth Social platform.
These series of images showed young women in “Swifties for Trump” T-shirts, along with one that portrayed Swift asking people to vote for the Republican presidential candidate, dressed as Uncle Sam. The images appear to be screenshots of blue-ticked right-wing X accounts that have previously been known to spread misinformation.
Swift has not publicly commented on this matter, but she had previously in the 2020 U.S. presidential election cycle openly criticized Trump over the country-wide protests that erupted after the police murder of George Floyd.
“After stoking the fires of white supremacy and racism your entire presidency, you have the nerve to feign moral superiority before threatening violence? ‘When the looting starts the shooting starts’??? We will vote you out in November. @realdonaldtrump,” she tweeted at the time.
Trump also shared an AI-generated image that showed Democratic candidate Kamala Harris holding a communist rally, as well as a deepfake video of him dancing with the X owner, Elon Musk, who has endorsed him.
Calls for Legislation Get Louder
Massimo Sterpi, the chair of AI Task Force of Union Internationale des Avocats, explained the likely rationale behind the latest deepfake from Trump. According to him “[Trump is] trying to take advantage of Swift’s gigantic group of fans.”
“As a Plan B, considering that Taylor Swift is very likely to publicly support Kamala Harris, that (fake) endorsement for Trump is also a way to ‘neutralize’ her (possible) future endorsement for Kamala Harris: in fact, at least part of the voters will be confused about whom was endorsed by Taylor Swift,” Sterpi said. “In short, either people will wrongly believe that Swift endorsed Trump or they will be uncertain about whom she was actually endorsing. Almost a win/win for Trump.” [sic]
It’s not just Trump but his supporters have also taken to deepfake AI-generated images since he first announced his nomination. His supporters earlier this year faked Trump’s image with people of color to target the African American voter base for the Republicans.
“Deepfakes are the new wild card in our elections.”
But Trump’s latest spread of misinformation has got even the experts worried. Dr. Chase Cunnigham, also know as Dr. Zero Trust, sounded the alarm and said, “Don’t care about the party or politics BS but influence can affect outcomes and deepfakes are the new wild card in our elections.”
Sterpi shared a similar view and urged legislators to take it seriously. “Legislation urgently needed to prevent these blatant abuses!”
Meanwhile, Tehran Behind Trump Hack and Influence Operations
Trump’s deepfake antics coincidentally took place on the same day the U.S. intelligence community called out Iran for its covert Influence operations.
Iran has escalated its cyberattacks targeting U.S. presidential campaigns in an effort to influence the upcoming election, according to top U.S. cybersecurity agencies. In a joint statement released Monday, the Office of the Director of National Intelligence (ODNI), the Federal Bureau of Investigation (FBI), and the Cybersecurity and Infrastructure Security Agency (CISA) linked a recent cyberattack on former President Donald Trump’s campaign to Iranian hackers.
“Iran seeks to stoke discord and undermine confidence in our democratic institutions,” the agencies said. “Iran has furthermore demonstrated a longstanding interest in exploiting societal tensions through various means, including through the use of cyber operations to attempt to gain access to sensitive information related to U.S. elections.”
The agencies observed that this election cycle has particularly seen an aggressive approach from Tehran. This includes influence operations aimed at the American public and direct cyber operations targeting presidential campaigns. The intelligence agencies reported that “the Iranians have through social engineering and other efforts sought access to individuals with direct access to the presidential campaigns of both political parties.”
“These efforts, which include data thefts and unauthorized disclosures, are intended to sway the U.S. electoral process,” the agencies said, noting that such tactics are not new and have been used by both Iran and Russia in the past, not only in the U.S. but also in other countries around the world.
This ramp up in influence operations targeting the current election cycle, featuring former President Trump and Vice President Kamala Harris, is very important to Iran’s national security interests. “The IC has previously reported that Iran perceives this year’s elections to be particularly consequential in terms of the impact they could have on its national security interests, increasing Tehran’s inclination to try to shape the outcome,” the agencies said.
The FBI is working closely with the affected campaigns to mitigate the impact of these cyberattacks. “Protecting the integrity of our elections from foreign influence or interference is our priority,” the agencies affirmed. “We will not tolerate foreign efforts to influence or interfere with our elections, including the targeting of American political campaigns.”
Microsoft, Google, OpenAI, Confirm Iran Attribution
This official attribution from the intelligence and cybersecurity agencies to Iran follows a recent cyberattack on the Trump campaign, an incident Trump has publicly blamed on Iran. In response, tech giants Microsoft and Google published reports detailing Iranian cyberattacks on high-ranking U.S. and Israeli officials, as well as on members of both presidential campaigns.
OpenAI, the parent company of ChatGPT – a large language model, concurred to Microsoft’s findings last Friday saying it had shut down a network of ChatGPT accounts involved in an Iranian influence operation.
The AI-generated content developed using ChatGPT targeted a range of topics, including the Gaza conflict, Israel’s participation in the Olympics, the U.S. presidential election, Venezuelan politics, and Scottish independence.
“This operation used ChatGPT for two purposes: generating long-form articles and shorter social media comments,” OpenAI said. However, its analytics into the performance of these AI-generated posts suggested no “meaningful” audience engagement.
Iran’s mission to the United Nations issued a statement calling the allegations “unsubstantiated and devoid of any standing.”
“As we have previously announced, the Islamic Republic of Iran harbors neither the intention nor the motive to interfere with the U.S. presidential election,” it added.
The intelligence and cybersecurity agencies called for increased resilience from online platforms on the back of these covert influence operations from U.S.’ adversaries. “Just as this activity demonstrates the Iranians’ increased intent to exploit our online platforms in support of their objectives, it also demonstrates the need to increase the resilience of those platforms.”