Experimental Evidence for Efficiency Gains on Trust via AI-Mediated Communication

The growing integration of AI tools in communication promises enhanced efficiency and productivity. However, concerns about the erosion of trust and authenticity have fueled debates over the disclosure and regulation ofAI-mediated communication. It remains unclear a) how AI mediation influences trust, b) whether the efficiency gains outweigh potential compromises in trust, and c) how proposed transparency policies would impact these trade-offs. To address these questions, we conducted two online studies (total N = 1637). We adapted the well-established incentivized trust game to measure objective behavioral trust. Additionally, we use subjective, self-reported trust ratings. Before each game, some participants could write a message with or without an AI assistant to elicit trust. We also systematically varied actual and expecteddisclosure of AI assistance. Results reveal that AI assistance had little to no consequences for trust, even when AI use was transparent. However, participants using AI assistance composed their messages more quickly, achieving higher returns of trust per writing time invested. Importantly, this efficiency boost was not suppressed by expected or actual disclosure of AI involvement. Contrary to prevailing concerns, these results suggest that AI-mediated communication could enhance efficiency and facilitate trust-building rather than undermine it.

osf.io/preprints/psyarxiv/tb37z

About
Latest Posts

Ryan Watkins

Professor at George Washington University

I am a Professor with Human-Technology Collaboration and Educational Technology programs at George Washington University in Washington DC. I have written 12 books and more than 100 articles, and I co-host of the Parsing Science podcast where scientists tell the stories behind their research. I am also the developer of the WeShareScience.com online platform for sharing research videos, and SciencePods.com where researchers can create free podcasts about their science. My research interests include human interactions with intelligent machines, needs, needs assessments, and instructional design.