As the United States prepares for the upcoming midterm elections next month, artificial intelligence (AI) technologies are playing an increasingly prominent role in shaping voter information and engagement across social media and search platforms. Researchers, political experts, and watchdog groups have raised concerns over the potential risks and inconsistent responses from tech companies regarding the accuracy, transparency, and oversight of AI-generated election content.
According to a recent survey by the American Association of Political Consultants Foundation, one in five voters plans to use AI to research candidates or ballot measures, while twice that number have already encountered election-related information summarized by AI tools. Despite this growing reliance, the survey found that only 1 percent of respondents consider AI to be their most trustworthy source for election details.
Social media platforms and search engines are now offering advanced AI features such as chatbots, multimedia content generation, and automated moderation services. These tools can provide overviews of political candidates’ policies or surface information about voting logistics like polling locations and deadlines. However, concerns persist that AI is also enabling the proliferation of misleading or false information. For example, automated systems have been linked to the amplification of biased narratives, fabricated images, and foreign influence campaigns, including state-sponsored disinformation tied to Russian entities.
Studies have revealed that numerous partisan websites—often referred to as “pink slime”—generate content primarily intended to train AI models rather than inform voters. Many of these sites produce material favorable to specific political outcomes, frequently targeting Democratic candidates in select House races. Experts warn that such networks compromise the integrity of AI systems by feeding them low-quality or deliberately distorted information, thereby affecting chatbot responses and search results that voters may encounter.
While tech companies have taken various measures to address these issues, critics argue that their efforts remain fragmented and insufficient. For instance, Google highlighted its use of AI to connect users with verified election resources and labels AI-generated content clearly, while Meta emphasized its real-time threat modeling and rigorous testing of AI outputs related to elections. Meanwhile, TikTok announced policies to flag AI-created content and empower users to add contextual notes on posts. Despite these initiatives, transparency about AI moderation processes and safeguards is limited.
Some AI developers have adjusted their models to reduce the risk of spreading false election narratives. OpenAI, for example, has shifted from generating realistic images that could be mistaken for evidence of voter fraud to producing cartoonish depictions to signal satire or commentary. Nevertheless, recent analyses indicate that other major AI providers may have loosened controls over election-related misinformation.
Lawmakers, including Senators Adam Schiff and Alex Padilla, have sought greater clarity from leading technology companies on their AI strategies to defend election integrity and combat disinformation. They expressed concern about the uneven and often opaque application of AI policies just weeks before voters begin casting ballots.
Experts emphasize that the rapid incorporation of AI into political discourse requires stronger human oversight and consistent standards to mitigate risks. As Deen Freelon, a communication professor specializing in digital politics, noted, the nation is navigating “uncharted territory” in how AI intersects with democratic processes. With AI tools deeply embedded in election information ecosystems, the balance between fostering innovation and safeguarding accurate, trustworthy content remains a critical challenge.