When people encounter a questionable political claim on social media, many no longer turn first to journalists, fact-checking organizations or even search engines.
Instead, they ask AI.
From ChatGPT to chatbot features embedded within social media platforms, artificial intelligence is rapidly becoming a new source of information verification. But as more people rely on AI to determine what is true and what is false online, an important question remains: How well do these tools actually work?
That question is at the center of research being conducted by Rice University doctoral student Agustin Prinetti, who will spend the fall semester in Paris studying how people use AI chatbots to fact-check political content and navigate misinformation.
Prinetti recently received the highly competitive Chateaubriand Fellowship from the Embassy of France in the United States, which supports outstanding doctoral students from American universities conducting research in France. The fellowship will allow him to continue his dissertation research as a visiting fellow at Sciences Po, one of Europe’s leading institutions for political and social science research.
“I’ll be focusing on my dissertation, which studies the growing use of AI chatbots to verify online posts about political events and identify fake news,” Prinetti said. “These new tools are increasingly available within social media platforms, and users across countries are increasingly turning to them to fact-check political content.”
The research arrives at a moment when concerns about misinformation, political polarization and trust in information sources continue to grow around the world. At the same time, AI companies are positioning chatbots as tools that can help users evaluate claims, summarize complex issues and answer questions about current events.
One of the central questions driving Prinetti’s research is why people are choosing AI over traditional sources of verification.
“My dissertation explores why more and more users rely on this tool even in regions like Europe, where there’s a high availability of fact-checks produced by journalists and experts, and under what conditions these tools work better or worse,” he said.
The answers could have implications far beyond academia.
As governments, technology companies, journalists and citizens grapple with the spread of misinformation, understanding how people interact with AI-powered fact-checking tools may help researchers identify both opportunities and risks. Insights from the work could inform future approaches to combating misinformation and improving public access to reliable information.
Over the past year, Prinetti has analyzed large volumes of social media data to examine how users from several countries, including France, engage with AI chatbots when evaluating political information. Support from Rice’s Center for Computational Insights on Inequality and Society and the Social Sciences Research Institute helped make that work possible.
The fellowship will allow him to move beyond the data and study those interactions firsthand.
“Now this opportunity gives me the chance to spend a semester in one of the settings I’ve been studying from my desk,” Prinetti said. “I’ll be able to interact directly with users, relevant actors and local researchers and expand the work I’ve done observationally.”
France offers an especially valuable setting because it allows him to compare how AI tools are adopted in a different political and media environment from the U.S.
While social media platforms often deploy the same technologies globally, countries vary in ways that can shape how those technologies are used and trusted. The availability of professional fact-checkers, levels of political polarization and exposure to foreign misinformation campaigns can all influence how people seek and evaluate information.
“The digital phenomena I study rarely stop at national borders,” Prinetti said. “Experiences like this help make the research more grounded and allow us to understand how these new tools, which are usually developed in the U.S., affect and are adopted by citizens living in different political and media environments.”
Prinetti will be based at Sciences Po’s Center for Political Research, where scholars study topics ranging from political trust and media consumption to the societal impacts of emerging technologies.
As AI increasingly shapes how people consume news and evaluate political information, understanding when these tools help users find reliable information and when they fall short will become increasingly important.
Through his work in France, Prinetti said he hopes to help answer that question, offering new insights into how citizens, researchers and policymakers can navigate an information ecosystem where AI is playing a growing role in determining what people believe.