The Impact of Generative AI on Spreading Election Misinformation
Generative AI technology, such as ChatGPT and Microsoft Copilot, has recently come under scrutiny for its role in spreading misinformation during the presidential debate between former President Donald Trump and President Joe Biden. Despite the claim about a “1-2 minute delay” in the CNN broadcast being debunked by CNN itself, both ChatGPT and Copilot confidently shared this false information with users.
One of the main concerns surrounding generative AI is its ability to confidently hallucinate information and scrape unverified real-time data from the web. This combination creates a perfect formula for spreading inaccuracies on a wide scale. As the U.S. presidential election approaches, fears about the impact of chatbots like ChatGPT and Copilot on voters are becoming more acute.
NBC News conducted an investigation into the issue and asked several AI chatbots, including ChatGPT and Copilot, as well as Google Gemini, Meta AI, and X’s Grok, about the alleged broadcast delay. Despite CNN’s debunking of the claim, both ChatGPT and Copilot continued to share this falsehood as fact in their responses. Copilot even cited a since debunked claim from former Fox News host Lou Dobbs’ website to support its incorrect response.
This incident is just one example of generative AI’s role in spreading election misinformation. A report by research company GroundTruthAI found that Google and OpenAI LLMs provided inaccurate responses 27 percent of the time. Additionally, AI Forensics and AlgorithmWatch discovered that Copilot gave incorrect answers about candidates and election dates, as well as hallucinated responses about Swiss and German elections.
The spread of misinformation through generative AI poses a significant threat to the integrity of elections and public discourse. It highlights the need for increased accountability and fact-checking measures when using these technologies. While generative AI has many potential benefits, it is crucial to address its limitations and ensure that it is not used as a tool for spreading false information.