AI could harm democracy. It’s also the tool that could save it
(NewsNation) — Artificial intelligence-fueled deepfakes and misinformation pose a real threat to democracy, but those on the front lines say AI could be the best weapon to fight against harmful content.
“There’s a lot at stake, everything’s at stake, democracy is at stake,” said Emily Spratt, a professor at the Insitute of World Politics and Columbia University specializing in how AI shapes culture. “We’ve got to get this right and make sure that there are no foreign influences or bad actors out there … by using technology to detect foul play of technology.”
The problem is not a new one — operatives for Iran and China made AI-generated content and audio and video deepfakes designed to influence the outcome of the U.S. 2020 election, CNN reported in May.
This year is particularly important, as roughly 2 billion people in about 60 countries — including India, the United Kingdom and the U.S. — are choosing their leaders in elections.
Combatting and recognizing the misuse of AI to create deceptive generative content “is a bipartisan issue,” Ashley O’Rourke, a member of Microsoft’s Campaign Success team, said during a forum at the Republican National Convention on Monday.
O’Rourke’s Campaign Success team at Microsoft is focused on helping fledgling campaigns harness AI for good and “evangelizing” the successful uses of technology.
O’Rourke and co-presenter Ginny Badanes explained campaigns can find effective uses of AI, including using the technology to draft communications, create the first pass on creative images, convert a stump speech into posts for social media and coalition building.
“We give them some best practices for getting started … do things like ‘how to write a prompt,’ ‘how to use (AI) safely,’ ‘think about how your data might be used when utilizing an AI tool,’” O’Rourke said.
Already in this election cycle, there have been notable deepfake scams targetting politicians on both sides of the aisle.
In January, thousands of New Hampshire voters were targeted by a deepfake robocall of President Joe Biden. The call was designed to sound like Biden and claimed voting in the state’s primary would preclude voters from casting a ballot in November.
The culprit, later identified as Democratic political consultant Steve Kramer, now faces a proposed $6 million fine by the Federal Communications Commission and several felony charges in New Hampshire.
Just last month, Republican Gov. Spencer Cox of Utah was the target of a deepfake video in which he was portrayed admitting to fraudulently gathering signatures for his gubernatorial campaign.
“I do think that more control and more guardrails on AI are necessary (to safeguard democracy),” Spratt said. “Regulation and policies are absolutely necessary. I don’t think that this is the type of thing that the market can just work itself out.”
Though there is no comprehensive federal legislation to combat the misuse of AI-generated content and deceptive use of deepfake technology, Congress is considering several bills that could institute more guardrails, and several U.S. states have passed laws or are considering policies to do the same.
AI deepfakes worse for women
Not only could the misuse of AI undermine the outcome of elections, it could widen an already stark gender gap in politics as more women are deterred from running for office, said O’Rourke and Badanes.
There are a record number of women in Congress in 2024, though that translates to just 25% of U.S. Senate seats and 29% of U.S. House seats, according to the Center for American Women and Politics.
Women are disproportionately targeted by abusive rhetoric and pornographic deepfakes, O’Rourke and Badanes said. They fear this practice could widen the representational divide that has closed over the decades.
Lauren Leader, the CEO and founder of the nonprofit group All In Together, said these threats often hold women back from participating in politics. The group is a nonprofit charged with encouraging multiracial, multigenerational women to participate in American civic life.
“So many women will say that the reason why they don’t want to do local politics, or they don’t want to get involved at any level, is because it feels so scary and dangerous,” Leader said.
She was a target of online harassment when she ran as a Democrat in what she described as the “heavily Republican town” of Harrison, New York. Leader was elected as the sole woman on the Town/Village Board in 2019.
“There was just a real relentless, relentless harassment online — targeting of my kids, posting of my address, and of my license plate on my car, and just a ton of disinformation,” she said. “And it is very frustrating because there’s really nothing you can do.”
Leader says her goal with All In Together is “to try to expand the impact and agency of women to make sure … that we have equal representation at all levels, and that our voices are being heard and we have a seat at every table.”
How AI can transform campaigns for good
As part of Monday’s forum, speakers shared ways they thought campaigns around the world have harnessed AI-generated content to productively continue political discourse.
During the forum, Badanes pointed to how the jailed ex-prime minister of Pakistan, Imran Khan, directed his staff to create deepfakes of himself to spread messages among his supporters.
That was not a deceptive practice, she said, because the public was aware that he was using technology to speak to supporters.
On a much smaller scale, O’Rourke pointed to the campaign of a candidate running for mayor in a small town who used AI to save time on some of the campaign work.
“If you’re at the federal level, you might have the resources to employ an entire team of copywriters, an entire digital team,” O’Rourke said. “But if you’re running at the local level and you also have a full-time job, you’re the one writing your fundraising appeals. And so, for instance, if a candidate is leveraging an AI tool to write her fundraising appeals and recreate new forms of it for her appeals, it helps save her time.”
