Existing interventions focus on de-bunking or pre-bunking, which are often ineffective against deeply held beliefs. Additionally, there is a lack of experimental evidence on feeling understood, as it has been studied correlationally, but rarely experimentally induced, and few studies have tested whether openness persists beyond the immediate interaction. Therefore, Huber (2025) strives to bridge these gaps by investigating whether feeling understood increases openness to opposing information, and under which conditions might this openness be sustained or resisted over time. 

Participants engaged in a 6-minute conversation with an AI chatbot before reading articles that challenged their perceptions of vaccinations and climate change. The conversations differed in that some were neutral, deflective, friendly, corrective, or understanding- focused. Huber assessed the participants’ perceived article credibility by their rated trustworthiness, accuracy, and bias with a five-item scale. 

• Understanding chat: validates participant’s views without correction
• Friendly: positive but unrelated to topic
• Corrective: provides factual corrections
• Neutral: expressed indifference 
• Deflective: ignores participant’s view

Further, to investigate whether the effects of feeling understood persisted beyond immediate interaction, the participants returned after 60 days and read a new political article without a chatbot interaction. 

Huber (2025) found that friendly and understanding- focused conversations increased feelings of understanding, and feeling understood correlated to enhanced perceived credibility of opposing information. However, although friendliness expressed short-term effects, possible resistance emerged later, while genuine understanding can initiate lasting cognitive and behavioral changes, not just temporary compliance.

Huber’s findings can translate into a classroom setting, where educators can emphasize the importance of ensuring that students feel understood before opening up to controversial topics. For example, they can acknowledge the students’ perspective and avoid immediate correction before presenting differing ideas. Additionally, this study illustrates the growth of AI, which can be used for cognitive warfare and potentially lessen societal polarization. 

To learn more about this study, visit the full research article linked here: 

https://www.sciencedirect.com/science/article/pii/S0747563225003176

This post is written by Eliza Hong.

Reference:

Huber, R. (2025). Countering division with friendliness: How feeling understood by a friendly AI triggers both openness and resistance. Computers in Human Behavior, 108870.