Deepfaked Democracy?
We are exploring the impact of generative AI on elections and getting excited about our upcoming event on “Media and January 6th”.
Bots & Ballots
Shannon McGregor, alongside UNC Center on Technology Policy’s Scott Babwah Brennen, appeared on WUNC’s “Due South” to discuss the uses of AI in elections, campaigns, and how it is regulated. Jeff Tiberii, co-host of Due South, begins with playing a clip of a robocall from “Joe Biden” where deepfaked Joe says, “It’s important to save your vote for the November election. Voting this Tuesday only enables the Republicans in their quest to elect Donald Trump again.”
Shannon discussed her concern about disinformation campaigns targeting electoral processes and how they involve clear voter suppression. Scott emphasized the lack of federal laws against voter suppression and false content, so the example of the Joe Biden robocall reinforces how important those laws are. Scott noted there have been a few bills proposed at the federal level without much movement. At the state level, however, there have been laws passed in 5 different states (California, Texas, Washington, Minnesota & Michigan) regarding generative AI and political ads. In January, a couple dozen more states introduced similar bills, as well.
Most major social media platforms, like Google and Meta, instituted policies in dealing with synthetic content and deep fakes. However, as Shannon points out, “the devil is in the application of those policies.” OpenAI, for example, says you can not use ChatGPT for political uses at all, but as Shannon noted “What is “political”? Is it just about elections? Is it just about campaigns?”
OpenAI, for example, says you can not use ChatGPT for political uses at all, but as Shannon noted “What is “political”? Is it just about elections? Is it just about campaigns?”
Jeff asks if there are opportunities for AI to be used as a tool to improve upon democracy and make elections more democratic, fair, and open. Shannon mentioned there is a possibility for new candidates on the scene to use generative AI in developing content, organizing mailing, and curating data lists of voters, and she also noted that using generative AI as a search tool is also a possibility in the future, but learning language models are only as good as the information that goes into training them.
So, what are best practices for individuals engaging on social media with disinformation (AI-generated or not)? Shannon noted that the people who might be most vulnerable to this kind of electoral disinformation are people who are new to the process of voting in general and/or new to the process of voting in North Carolina. They should talk to people that they trust about this— like people in their networks— and call their county state board of elections to check the veracity of information that they see. Shannon concluded, saying: “With generative AI, the same rule applies to what we would say about a lot of media content: if it makes you feel really good, or really bad, chill for a second before you share it or just dive into anything you see.”
With generative AI, the same rule applies to what we would say about a lot of media content: if it makes you feel really good, or really bad, chill for a second before you share it or just dive into anything you see.
Shannon also spoke with the Carolina Public Press about AI-generated messages, robocalls, and more in this election cycle: “The real danger around [use of texts and calls] is— whether it’s the campaigns or the people that support them— they deploy these because they’re trying to block people’s access to the ballot, since that’s the only way they can win, is if they block or limit access for certain types of people.”
Shannon noted that misdirection around the electoral process (ex. the dates of voting or the processes of voting) and how it tends to limit people’s access to the ballot is most concerning. First-time voters, those who aren’t frequent voters, or voters from out of state are most vulnerable to this misdirection, and Shannon recommends voters, “try to find a second source for the information that you receive about deadlines for things or how the electoral process operates.”
Media and January 6th
🚨📅Save the date!: CITAP is hosting an event featuring the co-editors and a number of authors of this volume on April 12th. More details coming soon… 👀🚨
Insurrectionists’ violent efforts to overturn the results of the 2020 election on January 6th, 2021, brought the fragility of American democracy to light. Over three years later, however, the national response to January 6th has largely failed to address legacies of racism, mis- and disinformation, and other threats to democracy fueling this attempted coup. We often take American democracy for granted, obscuring antidemocratic threats bubbling under the surface that allow events like those on January 6th to take place. Media and January 6th, a volume edited by Khadijah Costley White, Daniel Kreiss, Shannon C. McGregor, and Rebekah Tromble, brings together perspectives from experts across disciplines to help understand how the events of January 6th took place, how researchers should respond in light of these events, and what can be done to prevent similarly anti-democratic events from occurring in the future.
This edited volume emerged from a conference CITAP hosted on the one-year year anniversary of January 6th, 2021, “The Capital Coup One Year Later: How Research an Assess and Counter Threats to Democracy.” Attendees discussed the need to place January 6th in a broader historical context: contemporary political hostilities in the United States cannot be separated from unequal access to democratic participation based on race, gender, and other important aspects of identity throughout America’s history. Equally importantly, everyone—academics, politicians, and other members of the public—holds a unique social position impacting their interpretations of events and their ideas about how to address social and political problems. In times of emergency, then, we are obligated to use our expertise to protect democratic norms. As this volume’s editors assert,
“While some scholars believe that they should stay ‘above the fray,’ we believe that we are already firmly situated within it and therefore we have a set of responsibilities to clearly identify threats to and defend democratic institutions.”
In 22 chapters authored by researchers with expertise in political communications, media and journalism, rhetoric, sociology, and theory, Media and January 6th presents a variety of informed, thoughtful interpretations seeking to understand how the events of January 6th came about, how researchers across disciplines should respond, and how the American public broadly can work to prevent similar events from reoccurring. In bringing these perspectives together in one comprehensive volume, the editors take a vital step toward no longer ignoring the threats to democracy that the events of January 6th, 2021 made abundantly clear.
(Research Summary by Katherine Furl)
Publications and appearances
👏👏 👏 On February 14th, Data and Society (D&S) highlighted and thanked their advisors, who play a vital role in supporting D&S’s work, including Alice Marwick, “a Data & Society advisor and former fellow, she researches the social, political, and cultural implications of popular social media technologies.” 👏👏👏
Coming soon
March
March 20th @ 12pm in the Freedom Forum conference Center: UNC Sociology and CITAP are co-hosting speaker, Forrest Stuart. Lunch will be provided!
April
April 12th in the Freedom Forum Conference Center: CITAP will be hosting an event in celebration and reflection of the launch of the book “Media and January 6th” Stay tuned for more details!