Protecting Democracy in an Age of Deepfakes and Disinformation

In late January, a robocall went out to New Hampshire Democratic voters in a voice that sounded pretty close to President Biden’s. The message encouraged them not to vote in their state’s primary and to instead “save [their] vote” until November. The message was generated by a text-to-speech deepfake program and, thankfully, it wasn’t a very good fake. But it served as yet another sobering reminder of what voters in the United States and around the world are up against this year. Fears about the intersection of elections and disinformation abound. The news is replete with stories reminding us that information integrity, artificial intelligence (AI), and social media will be front and center during the world’s largest election year in history.

Over 4 billion people are expected to go to the polls this year against a backdrop that seems grim for democracies. A growing number of societies around the world are increasingly fractured and susceptible to dangerous identity politics. Political scientist Pratap Bhanu Mehta wrote recently that “nationalism is slowly strangling liberalism—a trend that could accelerate in a damaging way this year.” Authoritarianism is on the rise, and the consequences of political polarization are widely felt. Well-established technologies like social media, in combination with rapid advances in AI, appear to exacerbate (rather than cause) actual polarization and may well fuel a broad American perception that polarization is even worse than it is.

In societies characterized by low levels of trust, political division, and stark economic inequality, it’s easy for politicians and citizens to embrace extreme positions, assume that technology will inevitably erode democracy, or overemphasize finding tech solutions to problems with social roots. Instead, we need to get back to democratic fundamentals.

Technologies are human tools designed for human needs, and we do ourselves a disservice when we think of them as inscrutable collections of hardware and software that can only be governed by a narrow slice of society. How they are designed and what needs they serve matter. Technical and policy approaches to hard problems must be grounded in long-standing liberal democratic traditions that balance different societal interests, safeguard the rights of marginalized groups, protect a free press, and create safe environments for people to engage in open dialogue. Revitalizing these principles is more urgent than ever as Americans’ support for democratic norms erode.

In the world of tech policy, a focus on guardrails that protect democracy is vitally important. Tech governance challenges reflect the broader challenges democracies are confronting. Rapid developments in the digital age keep shifting the balance of power among governments, citizens, companies, and consumers. Our persistent focus on social media flows from an intuitive understanding that technological and democratic evolution is increasingly intertwined. Social media services depend on advertising business models that prioritize attention and don’t inherently privilege safety, privacy, or the veracity of content. This poses obvious challenges for democracy, even if the precise relationship among social media, mainstream media, political discourse, and election outcomes is unclear.

“No matter the tech policy question or challenge at hand, we must ground ourselves in time-tested democratic principles.”

AI adds another layer of complexity. This year, the tech policy community is tracking key issues that will arise repeatedly during this year’s electoral campaigns. These issues include generative AI’s impact on campaigns, social media companies’ decisions to downsize teams that track and counteract harmful speech, and the state’s role in fostering competition and innovation. Many countries have long been wrestling with how they can exert techno-nationalist control over the Internet and data flows. The United States is no exception.

There are causes for fear during this year of elections but also reasons for cautious optimism. Governments are engaged in a global AI governance race that can sometimes be counterproductive, but at least reflects appropriate attention on urgent challenges. In the United States, Congress is both focused on technology governance and yet constrained in what it can accomplish. But that’s not the case in the states, where supermajorities in 40 states ensure that legislatures will continue to be active laboratories of tech governance. And the executive branch has taken a people-centric approach to AI governance that is responsive to dogged advocacy from civil society groups. This approach is evident in the White House’s Blueprint for an AI Bill of Rights, a recent executive order on AI, and in the Office of Management and Budget’s implementing guidance. Meanwhile, U.S. officials at the national, state, and local levels are focused on safeguarding our elections from AI-driven disinformation and cyberattacks.

No matter the tech policy question or challenge at hand, we must ground ourselves in time-tested democratic principles. At the heart of every tough technology governance question is a social question fundamental to liberal democracies: How should we balance innovation, progress, and the need to ensure that the most vulnerable communities benefit most and are protected from disproportionate harms? These questions are central to the concepts of digital inclusion and equity that undergird New America’s approach to the relationship between technology and democracy.

Global elections will demand our attention in 2024. As countries confront challenges at the intersection of tech and democracy, voters will need to remind their elected leaders that healthy democracies prioritize their people in at least three ways. First, they promote robust civic engagement through voting and other means; they do not suppress it. Second, healthy democracies safeguard freedom of expression and zealously protect the freedom of the press. Third, strong democracies work hard to ensure that people are safe—online and offline—to engage civically and engage in open dialogue. No matter which technologies ultimately dominate the conversation in the 2024 elections and beyond, technocracy alone won’t illuminate the way forward. Tech policymaking must remain rooted in the fundamentals of liberal democracy to build an equitable and resilient digital future.

Follow The Thread! Subscribe to The Thread monthly newsletter to get the latest in policy, equity, and culture in your inbox the first Tuesday of each month.