The AI Arms Race in Political Campaigns


With the US presidential election looming later this year, many political commentators have expressed concerns about the use of AI in politics, particularly when it comes to election campaigns.

The rapid evolution of AI technology is already blurring the lines between real and fake content. Combined with many Americans’ distrust of the political establishment, this raises serious issues around how AI is used in politics and the potential impact on elections.

And this is not limited to the United States: in 2024, globally almost half of the world’s population will head to the polls, more voters than ever in history.

How AI is used in politics: influencing voter engagement

In the lead up to the US elections, AI is increasingly being used to help candidates reach a greater number of voters through phone conversations and chatbots – reducing the reliance on staff and volunteers for in-person voter engagement.

Democratic House candidate, Shamaine Daniels, recently launched an AI volunteer caller to speak to voters about what issues are important to them, alongside standard campaign strategies like digital advertising, direct mail and door-to-door campaigning.

New York City Mayor Eric Adams recently came under fire for using AI to make calls to city residents in a number of languages he doesn’t speak, including Yiddish, Mandarin and Spanish. While Adams claimed that the calls were a way to promote services to residents, the Surveillance Technology Oversight Project argues that this is both unethical and misleading.

Other organizations have voiced fears that AI could ramp up existing efforts to suppress voter turnout. UK think tank Demos has cited examples of ‘rumor bombing’ in key battleground states in America – bombarding voters with false social media messages about shootings or long queues at their local polling booths.

In 2020, far-right activists pleaded guilty to arranging thousands of robocalls intended to dissuade residents of minority neighborhoods from voting. There are concerns the possibility of using fake audio of politicians or public figures on robocalls could play a huge role in swaying voter engagement.

How AI is used in politics: data analytics and strategy formulation

AI offers enormous potential to analyze data relating to voting patterns and social media behaviors, in order to customize messages to different population demographics and regions. Political campaigns are now leveraging new technologies such as ChatGPT to produce campaign marketing materials, fundraising emails, and texts, and even the first drafts of candidate’s speeches.

Because AI’s data analytical capabilities enable far more precise audience targeting, campaigners can home in on swing voters – that is, people who have not yet decided which candidate they will vote for. With access to detailed personal data from social media, including the books they read and movies they watch, political campaigns are more easily able to craft their messaging around certain policies and partisan views.

This form of manipulation can become even more insidious. For example, voters may receive texts that appear to be from friends or a trusted source – but are, in fact, AI chatbots. As the conversation continues, the AI-generated persona gradually introduces information, tailored to the individual, which is aimed at swaying the voter to a particular political view.

Social media and deepfakes: navigating the new frontier

Deepfakes are AI-generated video, images and audio of events that did not actually happen. In the context of election campaigns, they can be used to sow deep confusion and mistrust among voters by creating entirely fictitious narratives.

A growing number of deepfakes of a political nature have already been widely circulated, impacting previous political campaigns. In Chicago’s 2023 mayoral primary election, a fake video surfaced of candidate Paul Vallas appearing to approve of police brutality. Vallas ultimately lost the race.

While running as a Republican presidential nominee, Ron DeSantis released an AI-generated video of his rival former President Trump apparently hugging White House medical advisor Dr Fauci without disclosing that the image was fake. AI-generated images of Trump being arrested also went viral in 2023, ahead of his indictment by a grand jury. More recently, a political ad published by the Republican National Committee depicts a dystopian AI-generated scenario under a second Biden term featuring explosions, waves of migrants, and martial law.

Current deepfakes can often be spotted by irregularities – such as extra limbs, strange eye movements, or audio not matching up with mouth movements. However, the technology is rapidly improving, making them harder to identify.

This places the onus on social media platforms to monitor and filter out deepfakes. Meta recently updated its rules to prohibit advertisers from using the company’s AI software to make political ads on Facebook and Instagram, and mandate disclosure of the use of third-party AI software. But is prohibition without sufficient moderation enough?

The future of political advertising

But it’s not all doom and gloom. In fact, some experts say that AI integration into the election space could lead to positive outcomes – if the technology is properly regulated.

For one thing, AI has the power to increase accessibility of information. Instead of reading policy positions listed on a website, individuals could interact with an AI chatbot that provides specific data-backed answers to their questions – helping them better understand the issues that affect them.

What’s more, historically disadvantaged communities – particularly those that speak English as a second language - could receive campaign messaging in their own language and dialect. The use of AI enhances this messaging by considering the translation as a complete text, rather than focusing on the accuracy of individual words.

Finally, one of the key challenges in election campaigns is the tendency to view voter blocks as a monolith with identical priorities and concerns – whether that is Black voters, Latinos, or suburban women. But AI can eliminate this narrative by harnessing powerful data analytics capabilities to create more effective campaigns that are both localized and personalized, forcing political candidates to consider the issues that voters genuinely care about.

FAQ: AI in politics 101

How are AI technologies specifically used to analyze voter sentiment?

AI tools analyze vast amounts of data from social media, polls, and other sources to gauge public opinion, enabling more targeted and effective campaign strategies.

What are the ethical concerns surrounding AI in elections?

The use of AI raises questions about privacy, the potential for manipulation through deepfakes, and the integrity of democratic processes.

What is the role of AI in government policy?

AI can play a critical role in policymaking by enabling governments to analyze large amounts of data to identify key patterns and trends, providing insights that will inform their decisions.