AI and Disinformation: When the Target Isn’t Readers, but Bots
May 13, 2025 — 08:30 am - 08:45 amTech Stage
Description
Why model training is now the new frontier in the fight against disinformation.
How is AI navigating the information ecosystem? What new risks is it introducing?
Chatbots are no longer just tools exploited by disinformation actors to generate realistic content quickly, at scale, and at zero cost. They are increasingly becoming a direct target for those spreading disinformation. The goal? To flood the web with false content in order to contaminate AI model responses.
But how do different bots perform? How prone are they to generating disinformation—or choosing not to respond—when asked about current events that have been the subject of fake news? And how does the accuracy of their answers change when prompts are submitted in different languages?
To break the cycle of disinformation, we must now focus on training the bots—and it’s crucial to understand how and why this must be done.
