Here's Why You Shouldn't Use AI to Vote (www.wired.com)

As Americans cast their votes in the midterms, many are unwittingly relying on AI chatbots for information about candidates and policies, with around 8 percent of surveyed adults admitting to using these tools for election guidance. However, concerns have emerged regarding the reliability of the information provided, with research from Forum AI revealing that nearly 40 percent of responses from popular chatbots like ChatGPT, Claude, and Google’s Gemini contained factual inaccuracies. Such errors include misrepresented voting deadlines and incorrect candidate information, underscoring the potential misinformation generated by these AI systems. This situation highlights a significant issue within the AI community: the dependency of chatbots on unreliable sources, often described as "pink slime" websites, which can skew the information they present. With a substantial portion of data sourced from politically motivated sites, chatbots can inadvertently propagate biases and misinformation. As noted by Nina Jankowicz, former disinformation czar, the gap between AI capabilities and public understanding poses risks to informed voting. This scenario calls for increased scrutiny of AI outputs and a reevaluation of how these tools are used in critical civic processes, emphasizing the need for improved training data and more robust verification mechanisms in AI-driven applications.