How to track what’s true as the war over disinformation in politics builds to a climax

Artificial intelligence is fueling an arms race between the purveyors of disinformation and those who are fighting it in this year’s high-stakes political campaign, but the best tool to defend against fake news is honest-to-goodness human intelligence.

That’s how two expert observers size up the escalating information war in the latest episode of the Fiction Science podcast.

“As a longtime AI researcher, I’ve become a huge fan of human intelligence,” says Oren Etzioni, the founder of Seattle-based TrueMedia.org, which uses AI to distinguish between genuine and faked imagery. “So, the first, second and third defense has to be media literacy and appropriate skepticism about what you see.”

Annalee Newitz, the author of “Stories Are Weapons: Psychological Warfare and the American Mind,” agrees that humans are “the most important part of the loop” in the fight against disinformation.

“We need technical tools. We need things like TrueMedia. We need access to APIs for social media platforms so that researchers can provide tools like TrueMedia for text and for posts that are mostly text-based,” Newitz says. “But ultimately, it is about people being wary of what they read that’s passed to them by any platform that they’re on, even if it’s something they hear from their neighbors.”

People have been hearing a lot of false narratives lately. For example, the claim that immigrants were “eating the cats … eating the dogs” in an Ohio community resulted in real-world harassment of the town’s residents. An AI-generated image of singer Taylor Swift seemingly endorsing GOP presidential candidate Donald Trump led her to come out and endorse Vice President Kamala Harris, Trump’s Democratic rival. And in the wake of Hurricane Helene, disinformation about the recovery effort sparked armed threats aimed at relief workers.

“Fake news, real consequences,” Etzioni says.

In “Stories Are Weapons,” Newitz documents how negatively framed storytelling has been used against Native Americans, against Black Americans, and against gays and lesbians. The dark lessons learned from earlier demonization campaigns are being applied to the stories we’re hearing during the current election cycle, Newitz says.

“One of the big focuses is immigrants, and the fear that immigrants are voting, the fear that immigrants are taking over our towns and doing things to our pets. But also, this is just a fear of people who are not like ‘us,’ right?” Newitz says. “The other group that’s being really targeted in this election cycle are trans people — particularly trans kids, but really just anyone who’s transgender or gender non-binary, like myself.”

AI is the latest frontier in the disinformation war: Clemson University researchers recently reported that a pro-Republican network of at least 686 AI-powered bots was programmed to respond to political posts on the X social-media network. And this week, U.S. intelligence officials said Russian operatives were behind an online smear campaign that used an AI-generated video to make false accusations of sexual misconduct against Democratic vice-presidential candidate Tim Walz.

The masters of disinformation are increasingly taking advantage of computerized data analysis tools to micro-target their stories. For example, HuffPost reported that a political action committee linked to Elon Musk targeted Muslim voters in Michigan with ads portraying Harris as a strong backer of Israel, while targeting Jewish voters in Pennsylvania with ads portraying her as pro-Palestinian and anti-Israel.

AI is also powering counter-efforts to defend against disinformation. TrueMedia.org is just one example. Just last week, a Seattle-based startup called Hiya rolled out Deepfake Voice Detector, a free browser extension that can flag AI-generated audio and video. Other Seattle startups targeting deepfakes include Loti, Certifi AI and Official AI.

Dozens of other organizations have joined the broader fight against fake news of all sorts. RAND, one of the nation’s oldest policy research centers, maintains a list of more than 80 online offerings designed to counter disinformation. There’s even an AI program called DebunkBot that gently coaxes conspiracy theorists to question their misguided beliefs. Scientists found that DebunkBot can change minds.

Not every battle has been won. Some of the organizations that kept tabs on online disinformation and issued alerts to social-media platforms and government officials have had to pull back, Newitz says, due to opposition from GOP lawmakers including Rep. Jim Jordan.

“Now that kind of research and that kind of tracking of misinformation is not available to us,” Newitz says.

Etzioni says he’s had to deal with similar concerns. “When I started TrueMedia.org with the goal of fighting disinformation, a number of people said to me, ‘Be careful — you’re going to be painting a target on your back for specifically Jim Jordan and others,’” he recalls. “I’m pleased to say that that’s not been the case, I think in part because we have focused on being nonpartisan.”

He notes that purveying fake news isn’t exclusively a right-wing pursuit.

“There is a tremendous amount of disinformation also being propagated by the far left, whether it’s having to do with Ukraine, with Gaza, with political candidates — you know, videos of Trump holding a gun,” Etzioni says. “Probably the most notable example is where Mark Ruffalo, the actor, shared with a million followers on Twitter an image of Trump with underage girls, supposedly on the Epstein plane. It was a fake image. It looked realistic to him, and he forwarded it.”

Is AI a net positive or a net negative when it comes to the disinformation war? “It pains me to say that in this context, it’s clearly a net negative,” says Etzioni, who served as the CEO of the Allen Institute for Artificial Intelligence before starting up TrueMedia.org.

That’s why Etzioni — and Newitz, for that matter — prioritize human intelligence over artificial intelligence. Here are a few of their recommended defense mechanisms:

• Take a pause: “When you have an emotional reaction to what you’re seeing, it’s a great time to take a moment and ask: Am I sure this is bona fide?” Etzioni says. “Am I sure this is real? Do I know where this came from?”
• Check the source: “Have multiple sources for your information,” Newitz says. “And when I say ‘multiple sources,’ I don’t mean, listen to somebody from the far left, and then listen to somebody from the far right, and try to find yourself in the middle. I mean, pay attention to newspapers, podcasts, radio news that actually have credentials and that may have dozens of years, if not a century, of track records, of being news sources that are trustworthy.”
• Pay attention to fake-news themes: “”If you’re starting to see, say, a rise in accusations about people voting who are not U.S. citizens, keep an eye on that,” Newitz says. “If you keep seeing the same myth again and again, that’s a red flag.”
• Keep receipts: “By ‘keep receipts,’ I mean, look to the past,” Newitz says. “Have you seen these kinds of stories circulating before? Are you hearing the same things about gay people that were said in the 1950s? Know that. Have that historical receipt.”

Both Etzioni and Newitz are deeply worried about what’s coming up. “My concern is particularly about the short term. Are there going to be bombshells? Attempts to sway this upcoming election that’s weeks away?” Etzioni says.

Newitz says, “I’ve got to confess, I am strapped in for a pretty scary election season. I really hope that I’m wrong.”

Both are more optimistic about the longer term. “We are definitely not doomed,” Etzioni says. “We’re resilient, and I believe it will take longer than it should. Certainly our legislative process is stymied, particularly at the federal level, but I do believe that we will figure it out.”

Newitz says one of their hopes for the future is that people on both sides of today’s political divide will find ways to tell better stories together, in person and online.

“Take a right-wing person and a left-wing person who both love Dungeons & Dragons, and just watch them talk about D&D for 20 minutes, you know, and say, ‘Wow, I really agree with you that the new rules have done the Cleric dirty,’” Newitz says. “What does it look like when people agree with each other about stuff? That’s what I mean when I’m thinking about telling better stories.”

---

Visit the Techsploitation website to learn more about Annalee Newitz and their writings — including two books that were the subjects of previous Fiction Science podcasts, “The Terraformers” and “Four Lost Cities.” To learn more about what Oren Etzioni is up to, check out TrueMedia.org and this article about the venture. GeekWire’s archives offer additional background on the subject of disinformation.

My co-host for the Fiction Science podcast is Dominica Phetteplace, an award-winning writer who is a graduate of the Clarion West Writers Workshop and lives in San Francisco. To learn more about Phetteplace, visit her website, DominicaPhetteplace.com.

Take a look at the original version of this item on Cosmic Log to get Newitz’s recommendations for further reading, and stay tuned for future episodes of the Fiction Science podcast via Apple, Spotify, Player.fm, Pocket Casts and Podchaser. If you like Fiction Science, please rate the podcast and subscribe to get alerts for future episodes.