FBI uncovers deceptive AI deepfakes in 2024 election's final hours
WASHINGTON — The competing claims flying across our screens have heated up in recent weeks, but we ain’t seen nothing yet.
Election experts and tech firms are bracing for a flood of artificial intelligence-fueled deepfakes, coupled with a torrent of more traditional mis- and disinformation efforts, in these final hours of the 2024 election.
“The most perilous moment will come, I think, 48 hours before the election,” Microsoft President Brad Smith warned the Senate Intelligence Committee at a hearing earlier this fall.
Shortly after that hearing, Congress took the entire month of October off without addressing the quietly building storm that we’re now in the throes of, including from foreign actors intent on sowing discord into American elections.
“People think it's campaign against campaign, not necessarily,” Sen. Amy Klobuchar (D-MN) told Raw Story. “It may well be foreign entities.”
With few federal guardrails in place to combat deepfakes and chatbot armies, in the last days of this election, federal officials are leaning on the private sector and states to fight AI-infused fakes, lies and misleading information. And experts are warning voters to be on guard and question everything they see online.
ALSO READ: Donald Trump believes he’s going to lose
“Platforms are gonna have to take it on themselves. The agencies are gonna have to get involved. Individual states will get involved,” Klobuchar said.
But most tech companies have relaxed their post-Jan 6 efforts to combat mis- and disinformation. Federal funding for cybersecurity protections in local elections come in the form of Help America Vote Act (or HAVA) grants, which went from $425 million in 2020 to only $55 million — shared amongst all 50 states and the District of Columbia — this election year.
That has members of both parties expecting a wild, bumpy ride this week.
The fakes are flying
Over the weekend, the FBI issued a warning about two fake videos circulating online; one of agents making arrests over ballot fraud and another falsely depicting the FBI saying they won’t investigate Second Gentleman Doug Emhoff, the husband of Vice President Kamala Harris, for interfering in the election.
“The FBI is aware of two videos falsely claiming to be from the FBI relating to election security, one stating the FBI has apprehended linked groups committing ballot fraud and a second relating to the Second Gentleman,” reads a statement the FBI released Saturday. “These videos are not authentic, are not from the FBI, and the depictions are false.”
After a fake President Joe Biden robocalled thousands of New Hampshire residents urging them not to vote in the state's primary, the Federal Communications Commission (FCC) made it illegal to use AI-generated fake voices in robocalls. The FCC action has been praised on Capitol Hill, even as lawmakers know it’s a drop in the proverbial bucket.
Without congressional action, there are few new tools officials have as voters are inundated with next-generation fakes — whether it’s a believable deepfake video or AI working in the background to amplify conspiracies on social media — heading into Tuesday.
But in 2024, deepfake technology is melding with traditional disinformation efforts.
In Georgia, Secretary of State Brad Raffensperger has also been combating lies spread by Republicans like Rep. Marjorie Taylor Greene (R-GA) that voting machines are flipping people’s votes — “That claim was a lie in 2020 and it’s a lie now,” Raffensperger said — or another falsehood viewed by hundreds of thousands of people falsely claiming Haitian asylum seekers are sent voter registration packets.
This domestic misinformation environment has made this year’s election ripe for foreign meddling.
In September, the Justice Department charged employees of Russian state-controlled outlet RT with foreign election interference, alleging they dropped $10 million on unwitting right-wing influencers who in turn spread Kremlin lies, misinformation and propaganda in roughly 2,000 YouTube videos that were viewed more than 16 million times.
Then, a week later, the Treasury Department accused Russian nonprofit Autonomous Non-Profit Organization (ANO) Dialog and ANO Dialog Regions of using "deep fake content to develop Russian disinformation campaigns," which included "fake online posts on popular social media accounts …. that would be composed of counterfeit documents, among other material, in order to elicit an emotional response from audiences."
Other countries are using AI chatbots to influence the election. China has been actively targeting tough-on-China candidates in "down-ballot races,” according to Senate Intelligence Committee Chair Mark Warner.
Iran’s been sophisticated and precise in its interference, using AI to help construct realistic fake websites that are hyper-targeted towards specific communities, including its “Not Our War” site aimed at veterans and active duty soldiers, “Afro Majority” aimed at Black communities and “Savannah Time” aimed at conservatives in the battleground state of Georgia.
In these waning hours of the 2024 election, experts expect a chorus of AI-powered chatbots to drum up fear, anger and distrust, while also rolling out fake videos and manufactured audio hyper-targeted at vulnerable voters.
Legislative measures to label deepfakes, increase transparency in AI and protect the work and likenesses of journalists, artists and actors from being used to generate AI content may have scared off some bad actors, but without congressional action, experts are bracing for a flood of fakes from foreign and domestic actors alike.
“Apparently we're just gonna…hope for the best”
“Whatever happened to AI and deepfakes?” Raw Story asked earlier this fall.
“Apparently we're just gonna, you know, hope for the best,” Sen. Josh Hawley (R-MO) — one of the louder voices calling for AI guardrails in this Congress — told Raw Story. “I think it's dangerous.”
Hawley has teamed up with Klobuchar on efforts to combat deepfakes. He’s been frustrated that their bipartisan efforts haven’t seen the light of day in the Senate, even after much fanfare was given to artificial intelligence.
Last year, Senate Majority Leader Chuck Schumer and a bipartisan group of senators hosted nine, closed-door AI Insight Forums, but since then the Senate’s only been in Washington roughly three days a week and election-focused AI measures never hit the Senate floor.
“Do you blame Schumer for not devoting floor time to it?” Raw Story asked.
“Yeah,” Hawley replied. “But I think it’s probably leadership, in general.”
“Both sides?” Raw Story asked.
“I think so,” Hawley replied, including Senate Minority Leader Mitch McConnell in his critique of the stalled measures.
It’s not that simple, according to Schumer’s AI ally, Sen. Mike Rounds (R-SD) who helped spearhead last year’s private AI tutorials in the Senate.
Instead of a sweeping artificial intelligence bill, Rounds says to expect AI measures in every major measure that comes out of Congress from now on.
“It's not like there's going to be a megabill coming,” Rounds told Raw Story. “I don't think there's going to be a major piece of legislation that's probably going to pass the Senate in the future without having something to do with modifications to address AI issues.”
For example, Rounds points to artificial intelligence measures tucked into annual spending — or appropriations — bills (like efforts to boost domestic AI research), the National Defense Authorization Act (or NDAA, which includes provisions to better define AI warfare), the reauthorization of the Farm Bill (where lawmakers are promoting AI technology in farming) and a myriad of other measures winding through Congress.
But those measures are for another day. True to form, ahead of the election Congress is out this week, even as generative AI is spinning new deepfakes in real-time that are then being weaponized against vulnerable communities and susceptible — or lazy — influencers alike.
“Are we ready for it?”
“Are we ready for it?” Raw Story asked.
“Were we ready for it in 2016? Were we ready for it in 2018? Were we ready for it in 2020?” Rounds asked. “You get better at it. In 2018, we really knocked down a ton of stuff, but we were able to do a lot of that through our cyber activities. We've gotten better with our cyber stuff, but we've got adversaries that are getting better at influencing.”
Much of it comes down to self-policing the information and disinformation we consume. Rounds says it’s buyer beware.
“If you're going to listen to information of any kind, you know, check and see where the source is, check and see whether there's credibility on those sources. And sometimes the American people, they don't take the time to do that,” Rounds said. “Particularly when they're inundated with multiple social media sources that they think must be more accurate than what they really are.”
Rounds and others say bad actors don’t get a pass on U.S. laws just because they’re deploying new technologies in the election.
“The foreign interference that we're seeing is illegal, regardless of if it includes AI or not, so having another tool available to an adversary does not change or make legal interference activity on their behalf,” Rounds said.
That’s little comfort to members of both parties who’ve been watching foreign interference efforts dupe Americans this cycle.
“Thomas Jefferson used to say, among other things, he used to say, when the American people know the truth, they won't make a mistake. In the situation we're in today, it can be folks just don't know what the truth is, ” Sen. Tom Carper (D-DE) told Raw Story. “But to have the ability, through AI and other mechanisms, can make the job of democracy even tougher.”
While local and federal officials alike are braced for the worst, others say politicians have always adapted to new mediums. No matter how revolutionary generative AI proves to be, the argument goes, everything has remained the same about American elections in 2024.
“It's going to be a gut check”
In battleground states, lawmakers like Sen. John Fetterman (D-PA) aren’t paying much attention to foreign interference — whether AI-assisted or not — in this election.
“They might try to do that kind of s---,” Fetterman told Raw Story this fall. “That's an issue, but it's not going to be the defining kind of issue.”
That’s in part because it’s become hard to impossible to differentiate foreign falsehoods from those being spread regularly by Trump himself — from him claiming he won the 2020 election (which he lost to President Biden) or him pushing conspiracies about pet eating in Springfield, Ohio (which has been thoroughly debunked).
Then there’s social media. In past elections, many tech firms worked to actively combat mis- and disinformation, but in 2024, X owner Elon Musk — whose tweets are boosted beyond his 202 million followers these days — has become one of the nation’s most notorious spreaders of actual fake news.
A September analysis of 171 of his posts over a 5 day period this fall by the New York Times showed roughly a third of Musk’s posts “were false, misleading or missing vital context.” Musk’s new AI software, Grok — a Wikipedia-like function built into X — is unreliable and has become a regular spreader of misinformation, as it’s transformed conspiracies of ‘rigged’ elections into reality for the tens of millions on the social media platform.
Then there’s Robert F. Kennedy Jr., hailed as a rockstar at Trump rallies. But he’s a notorious anti-vaxxer who is now promising to remove fluoride from America’s water if Trump gives him the reins of the Department of Health and Human Services (HHS), as the former president has promised.
In a mis- and disinformation climate like this, it’s no wonder voters are confused with where reality ends and rhetoric begins, let alone all the deepfakes in between.
But even with all the new tech, those in battleground states say, as always, the fundamentals will decide this year’s election.
“I wish I could say something more sexier or interesting or shocking, but it's going to be ridiculously close,” Fetterman said. “It's not about polls, it's not about money, it's not about, you know, debates or endorsements — it's about the choice, and it's going to be a gut check.”
NOW READ: To hell with Trump's voters
Democratic presidential nominee and US Vice President Kamala Harris speaks during a campaign rally at Michigan State University in East Lansing, Michigan, November 3, 2024. © Carlos Osorio, Reuters
Republican presidential nominee and former US President Donald Trump walks onstage during a rally at Atrium Health Amphitheater in Macon, Georgia, November 3, 2024. © Brian Snyder, Reuters