As the United States gears up for its high-stakes presidential election on Nov. 5, fears have escalated about the potential of technology to disrupt the vote. Concerns about social media and its role in spreading misinformation—spurred on by Washington’s adversaries—have been in the public conversation for a decade, but officials warn that interference efforts by China, Iran, and Russia have kicked into overdrive around the 2024 election.

One of the biggest fears is the impact of artificial intelligence, particularly large language models—such as OpenAI’s GPT-4—which can generate text, audio, and video on demand and threaten to turbocharge misinformation. While there has been no conclusive evidence that any of the dozens of elections held so far this year have been swayed by AI-generated misinformation, scientist and AI expert Gary Marcus says it’s a disaster waiting to happen.

“I am certain that we will see more AI-generated misinformation in the upcoming U.S. election,” Marcus told Foreign Policy. Troubling early examples in this year’s election cycle include a fake robocall in January that used President Joe Biden’s voice to discourage voters from going to the polls during the New Hampshire Democratic primary and a fake video that officials attributed to Russia last week, which showed a poll worker in Pennsylvania destroying ballots.

“I might be willing to accept that we have seen less of it than I anticipated. But we’re seeing some of it, and it only takes one close election to be flipped by this to really undermine the nature of democracy,” Marcus said. “I think it would be foolhardy to say: ‘Well, there’s been no major disaster yet, so we’re okay here.’ That’d be like saying we made a bunch of steamships, so this one’s invincible, and whoops, you hit an iceberg.”

Marcus said that this year’s election is part of what motivated him to very quickly write and publish his new book, Taming Silicon Valley: How We Can Ensure that AI Works for Us. “I wanted to get people to realize, as those elections take place, that AI policy is really, really important,” he said. “We screwed up social media, and we can’t afford to do the same with AI.”

Marcus’s book lays out the current state of artificial intelligence—particularly focusing on the generative AI that powers large language models—and the biggest risks it poses to society. He spends a lot of time warning about Big Tech’s influence over politics as well as the perils of government giving entrenched players such as Google, Meta, Microsoft, and OpenAI too much leeway without reining them in.

There is more of a public conversation around AI guardrails today than there was in the early days of social media, when policymakers failed to recognize the technology’s harms.

Last year, Marcus testified in Congress alongside OpenAI CEO Sam Altman, with both highlighting the risks of the technology and Altman stating that he welcomed regulation and independent oversight—though Marcus later said that Altman’s subsequent behavior and OpenAI’s lobbying against European AI regulation indicate otherwise. (“It’s becoming increasingly clear that the kind of things that [Altman] said when we stood side by side in the Senate are not entirely accurate,” Marcus told Foreign Policy.)

Taming Silicon Valley: How We Can Ensure That AI Works for Us, Gary Marcus, the MIT Press, 240 pp., $18.95, September 2024; The Tech Coup: How to Save Democracy from Silicon Valley, Marietje Schaake, Princeton University Press, 336 pp., $27.95, September 2024.

But Marcus said that policymakers are still in danger of making the same mistakes.

“You can’t allow regulatory capture where the big companies make the rules. You have to monitor the situation, and bad practices can become both culturally and legally entrenched,” he said. His book lists a series of potential fixes to problems with AI development and regulation, including privacy protections, greater AI literacy, and a strong AI governance agency both in the United States and globally.

“I think Washington is still in thrall with Silicon Valley and takes Silicon Valley at face value,” Marcus added. “Washington has to wake up and realize that in fact, Silicon Valley is in the midst of a huge power grab.”

That power grab is the emphasis of another book published in September: Marietje Schaake’s The Tech Coup: How to Save Democracy from Silicon Valley. Drawing on her decade of experience as a member of the European Parliament until 2019 and her post-political career as a fellow at Stanford University’s Cyber Policy Center and the Institute for Human-Centered AI, Schaake argues that the rapid digitization of our lives has left private companies as purveyors of rights and services traditionally guaranteed by governments, but without the requisite accountability.

Governments, she writes, have willingly “outsourced” those key functions to a handful of powerful CEOs. Examples abound through the book: Amazon oversees more than one-third of the world’s cloud computing capability, Palantir is the provider of choice of at least a dozen governments’ data systems, and Meta’s web of social media platforms has had a profound influence over global speech.

“The outsized power of tech companies … hurts democracy,” Schaake said in an interview with Foreign Policy. “This perspective is too often lost in an avalanche of incidents that are all very important and also urgent to solve, but that sort of cloud the view of the bigger picture. I wanted to write down that bigger picture challenge instead of talking about disinformation or cyberattacks or election interference,” she added.

It’s not just the tech giants, either. Schaake points out that the critical technology infrastructure that keeps our daily lives running, from undersea cables to data centers to semiconductor chips, is controlled by private companies—not to mention spyware and facial recognition software.

Although she welcomes the antitrust conversations around tech monopolies that have gained momentum on both sides of the Atlantic in recent years, she believes that a better way for governments to deal with these companies would be to think about the harms that they pose to democracy itself rather than just economic harms.

“Spyware companies are incredibly powerful and impactful and anti-democratic, but they would not meet the threshold of antitrust,” Schaake explained. “By identifying them as having significant societal power, there should be interventions possible that currently don’t really exist under one roof.”

There are few better exemplars of Schaake’s fears than the world’s wealthiest man. Elon Musk has a level of geopolitical influence that even his ultra-rich peers cannot claim: Both Ukraine’s war effort against Russia and the U.S. space program are increasingly dependent on his company SpaceX, and he has deployed the $44 billion algorithmic megaphone that he purchased in the form of Twitter (now X) to try to put former President Donald Trump back in the White House. But Schaake thinks that it is a mistake to focus on Musk or any other single tech billionaire.

“Elon Musk looks unique, but he’s not,” she said. “Private actors in the digital realm hold too much unaccounted power, and that is harmful for democracy writ large.”

There are signs that governments are pushing back against tech power—including against Musk himself. Brazil briefly banned X earlier this year over Musk’s refusal to block accounts accused of spreading misinformation in the country. And that move came shortly after Telegram founder Pavel Durov was arrested in France for not doing enough to curb illegal content on the platform.

Yet those remain isolated examples, and much of the global political will to act could be determined by the outcome of next week’s election. Trump has hinted he could put Musk in charge of a “government efficiency” body that would give the billionaire explicit political power, and Republican vice presidential candidate J.D. Vance has suggested pulling out of NATO if Europe regulates Musk’s companies.

“This kind of leveraging back and forth between U.S. geopolitical power and the private power of Musk, and presumably others, is unheard of,” Schaake said. “With [Democratic candidate and Vice President] Kamala Harris, we don’t know as much—I hope she will see this really as a key topic for everything that the Democratic Party and she herself should care about to really rein in the outsized power of tech companies.”

Regardless of the election results, Schaake said that the world needs a broader rethink of the relationship between governments and technology companies.

“My criticism of democratic governments is that they have not used their legitimacy and their primacy to create a comprehensive—or even noncomprehensive—framework around governing technologies,” she said. “If authorities choose to slam their fists on the table, they [can] still hold a lot of power.”