How AI Chatbots Are Shaping Voter Decisions in the Midterms
As Americans head to the polls today, a significant number will turn to artificial intelligence-driven chatbots for insights on candidates and policies. According to recent surveys by Pew Research, around 50 percent of US adults have interacted with an AI chatbot at some point, and 8 percent used these tools specifically for gathering election-related information ahead of the midterms.
The Rise of AI in Voter Decision-Making
The increasing reliance on AI chatbots for electoral information highlights a broader trend of technology integration in our daily lives. Chatbots, powered by advanced algorithms, promise to help users navigate complex political landscapes by providing concise, informative answers. However, this convenience comes with its own set of challenges.
Accuracy Issues with AI Chatbots
One of the most pressing concerns is the accuracy of the information these chatbots provide. Recent data from Forum AI reveals that nearly 40 percent of election-related answers given by major chatbots contained factual errors. For instance, when researchers inquired about Congressman Don Davis from North Carolina, the AI conflated him with a “former school superintendent” of the same name. Such inaccuracies could mislead voters, affecting their choices at the polls.
Outdated Information and Hallucinations
In addition to factual inaccuracies, much of the data offered by AI chatbots is often outdated or “hallucinated.” For example, last week, Google’s AI Mode refused to disclose the results of the 2020 elections, while its Gemini chatbot mistakenly indicated that the deadline to request a mail-in ballot in Pennsylvania was November 3rd, when in fact, it was October 27th. These errors highlight the limitations of AI in providing real-time, reliable electoral information.
The Problem with “Pink Slime” Websites
Further complicating the situation is the source material that AI chatbots rely on for training. Research by NewsGuard has found that nearly 50 percent of the time, chatbots reference what are termed “pink slime” websites—platforms that masquerade as reputable news sources but are funded by political or ideological factions. This not only taints the quality of information presented but also risks spreading misinformation during critical electoral periods.
More Read
Bot Behavior: Who Cites What?
Among the mainstream chatbots evaluated, OpenAI’s ChatGPT stands out as the most problematic, referencing pink slime sites in over 70 percent of inquiries related to candidates and policies. In contrast, xAI’s Grok chatbot fared better, citing these sites in only 30 percent of cases. This disparity raises important questions about the training data used in developing these AI models and their subsequent reliability in crucial conversations like those surrounding an election.
A Call for Caution
Nina Jankowicz, the executive director of the American Sunlight Project and a former disinformation czar during the Biden administration, warns us of the risks involved. “I don’t think that most Americans are prepared to navigate an election in the age of consumer AI,” she notes. As AI technology continues to evolve, she emphasizes that the gap between public expectation and AI capability must be addressed. “These AI models have been overpromised on what they can deliver, often leading to confusion and misinformation.”
The Impact on Voter Behavior
As more people turn to AI for electoral insights, it’s critical to consider how this reliance influences actual voting behavior. The simplicity and immediacy of AI-generated responses can sway undecided voters but also risk creating a skewed understanding of the political landscape. Encouragingly, this situation presents an opportunity for education about the limitations of AI, helping voters cultivate healthier habits when it comes to sourcing information.
Navigating the AI Landscape
For voters, being aware and discerning when using AI chatbots is more crucial than ever. Verification of information and cross-referencing with trusted sources can mitigate the risks associated with AI-powered advice. As these technologies continue to penetrate everyday life, making informed decisions based on accurate data will be paramount for active citizenship in a democratic society.
By understanding the strengths and weaknesses of AI chatbots, voters can better navigate their electoral experiences, ensuring that their choices are grounded in reliable and current information.
Inspired by: Source
