In the 2024 elections, artificial intelligence played a significant role, yet deepfakes were merely one aspect of its broader influence.
This year, 2024, marked the most significant election period in human history, with 3.7 billion eligible voters in 72 countries casting their votes. This was also the first election season where AI was extensively utilized, leading to concerns about deepfakes and AI-generated misinformation potentially sabotaging democratic processes. As we approach the end of 2024, it's interesting to examine how democracy fared during this time.
In a Pew survey of American voters this fall, nearly eight times as many respondents predicted AI would be used primarily for negative purposes in the 2024 elections compared to those who thought it would be used primarily for positive purposes. While there are genuine concerns and risks associated with using AI in political pursuits, its impact has not been wholly detrimental.
The infamous "death of truth" has not materialized, at least not due to AI. Furthermore, politicians are increasingly employing AI in constructive ways, provided it's used responsibly. However, given that these advancements often take place behind closed doors, the general public may not be privy to the full picture.
Engaging voters
One of the most impressive and beneficial uses of AI has been language translation, with political figures worldwide, such as Indian Prime Minister Narendra Modi and New York City Mayor Eric Adams, utilizing AI to translate their speeches and meetings for their diverse constituents.
Even when politicians themselves aren't speaking through AI, their constituents might be engaging with it. Google, for instance, introduced free translation services for an additional 110 languages this summer, offering real-time translation to billions of people through their smartphones.
Candidates have also taken advantage of AI's conversational capabilities to engage with voters. U.S. politicians Asa Hutchinson, Dean Phillips and Francis Suarez deployed chatbots of themselves in their presidential primary campaigns. The fringe candidate Jason Palmer even managed to outperform Joe Biden in the American Samoan primary, partially thanks to using AI-generated emails, texts, audio, and video. Imran Khan, Pakistan's former prime minister, also utilized an AI clone of his voice to deliver speeches from prison.
The most notable use of this technology was in Japan, where an obscure Tokyo gubernatorial candidate, Takahiro Anno, employed an AI avatar to respond to 8,600 questions from voters, enabling him to secure a fifth-place finish in a highly competitive field of 56 candidates.
An AI persona, named "AI Steve," even ran for office in the 2024 U.K. election.
Technical aspects
AI has also been utilized in political fundraising, with companies like Quiller and Tech for Campaigns marketing AI to help draft fundraising emails. Other AI systems aid candidates in targeting specific donors with personalized messages. Despite the difficulties in measuring the impact of these tools, political consultants remain interested in incorporating AI in fundraising efforts.
Polling has relied heavily on mathematics for decades, with pollsters continually incorporating new technologies into their processes. Techniques range from employing AI to extract sentiment from social media platforms (known as "social listening") to creating "synthetic voters" capable of answering tens of thousands of questions. While the effectiveness of AI applications in polling and strategic insights is still unclear, research is promising in light of the increasing challenge of reaching and surveying real humans.
On the political organizing front, AI assistants are used for an array of tasks, including crafting political messages and strategy, generating ads, drafting speeches, and coordinating canvassing and get-out-the-vote efforts. In Argentina's 2023 election, both major presidential candidates utilized AI to develop campaign posters, videos, and other materials. This trend was almost certainly replicated in various elections around the world in 2024, including the U.S., with a Georgia politician employing AI to produce blog posts, campaign images, and podcasts.
Misinformation and Deepfakes
AI-generated misinformation and propaganda did emerge, although not as catastrophic as initially feared. In a Slovakian election in 2023, fake audio discussing election manipulation went viral, but its impact remains unclear. In the U.S. presidential election, there was much hype around a robocall featuring a fake Joe Biden voice discouraging voters from participating in the Democratic primary, but it had minimal impact on the vote's outcome.
AI-generated images from disaster areas and viral deepfakes misrepresenting candidates' actions were similarly ineffective. Russian intelligence services attempted to influence U.S. voters using AI, but its success remains uncertain.
AI likewise contributed to safeguarding the data environment. OpenAI utilized its own AI models to hinder an Iranian foreign impact activity that aimed to spread discord ahead of the U.S. presidential election. Although anyone can utilize AI devices right now to produce lead-believable phony sound, pictures, and text, and this capability will stay, tech stages additionally utilize AI to mechanically manage content like hate speech and bigotry. This is a positive utilization case, improving content management more effective and sparing humans from examining the most serious offenses, yet there's space for it to turn out to be more adequate, more clear, and more reasonable.
There is plausibility for AI models to be significantly more adaptable and appropriate to additional dialects and nations than associations of human moderators. However, the realizations to date on stages like Meta show that a considerable measure more work needs to be done to make these frameworks reasonable and successful.
One thing that didn't make a great deal of difference in 2024 was AI designers' limitations on utilizing their devices for governmental issues. Regardless of market pioneer OpenAI's accentuation on prohibiting political uses and its utilization of AI to mechanically dismiss a large portion of a million solicitations to produce pictures of political competitors, the organization's execution has been disappointing and actual use is ubiquitous.
The genie is out
All of these patterns – both great and terrible – are probably going to continue. As AI gets more effective and capable, it is probably going to invade every part of governmental issues. This will happen regardless of whether the AI's execution is superhuman or suboptimal, regardless of whether it makes blunders or not, and regardless of whether the harmony of its use is positive or negative. All it takes is for one party, one crusade, one outside gathering, or even an individual to see an advantage in automation.
Bruce Schneier, Adjunct Lecturer in Public Policy, Harvard Kennedy School and Nathan Sanders, Affiliate, Berkman Klein Center for Internet & Society, Harvard University
This article is republished from The Conversation under a Creative Commons license. Read the original article.
In the future, the integration of AI technology in politics is expected to become even more prevalent, with politicians and candidates leveraging it for various purposes. This could include using AI for language translation, engaging with voters through chatbots, and even having AI avatars run for office.
For instance, the use of AI in political fundraising is already being explored by companies like Quiller and Tech for Campaigns, which aim to help candidates craft fundraising emails and target specific donors with personalized messages. Similarly, AI is being used in polling to extract sentiment from social media platforms and create synthetic voters that can answer tens of thousands of questions.
However, the use of AI also poses potential risks, such as the creation and spread of deepfakes and AI-generated misinformation. While AI was extensively utilized during the 2024 election period, concerns about its impact on democratic processes emerged. For instance, an infamous "death of truth" due to AI did not materialize, but its utilization often took place behind closed doors, leaving the general public unaware of its full impact.
As AI advances, it is crucial to ensure its responsible use in politics, striking a balance between its benefits and potential risks. This may involve improving AI frameworks to detect and manage content like hate speech and bigotry, as well as educating the public about the role of AI in politics and its implications for democratic processes.
