Lost Before Translation: Social Information Transmission and Survival in AI-AI Communication
arXiv:2602.17674v1 Announce Type: cross Abstract: When AI systems summarize and relay information, they inevitably transform it. But how? We introduce an experimental paradigm based on the telephone game to study what happens when AI talks to AI. Across five studies tracking content through AI transmission chains, we find three consistent patterns. The first is convergence, where texts differing in certainty, emotional intensity, and perspectival balance collapse toward a shared default of moderate confidence, muted affect, and analytical structure. The second is selective survival, where narrative anchors persist while the texture of evidence, hedges, quotes, and attributions is stripped away. The third is competitive filtering, where strong arguments survive while weaker but valid considerations disappear when multiple viewpoints coexist. In downstream experiments, human participants rated AI-transmitted content as more credible and polished. Importantly, however, humans also showed
arXiv:2602.17674v1 Announce Type: cross Abstract: When AI systems summarize and relay information, they inevitably transform it. But how? We introduce an experimental paradigm based on the telephone game to study what happens when AI talks to AI. Across five studies tracking content through AI transmission chains, we find three consistent patterns. The first is convergence, where texts differing in certainty, emotional intensity, and perspectival balance collapse toward a shared default of moderate confidence, muted affect, and analytical structure. The second is selective survival, where narrative anchors persist while the texture of evidence, hedges, quotes, and attributions is stripped away. The third is competitive filtering, where strong arguments survive while weaker but valid considerations disappear when multiple viewpoints coexist. In downstream experiments, human participants rated AI-transmitted content as more credible and polished. Importantly, however, humans also showed degraded factual recall, reduced perception of balance, and diminished emotional resonance. We show that the properties that make AI-mediated content appear authoritative may systematically erode the cognitive and affective diversity on which informed judgment depends.
Executive Summary
The article explores the transformation of information when AI systems communicate with each other, using a telephone game paradigm. The findings reveal three consistent patterns: convergence towards a neutral tone, selective survival of narrative anchors, and competitive filtering of arguments. Human participants perceived AI-transmitted content as credible but showed degraded factual recall and emotional resonance. The study highlights the potential erosion of cognitive and affective diversity in AI-mediated content, raising concerns about informed judgment and decision-making.
Key Points
- ▸ Convergence of AI-transmitted content towards a neutral tone
- ▸ Selective survival of narrative anchors over detailed evidence
- ▸ Competitive filtering of arguments, favoring stronger ones over weaker but valid considerations
Merits
Novel Experimental Paradigm
The use of a telephone game paradigm to study AI-AI communication provides a unique insight into the transformation of information.
Comprehensive Findings
The study reveals consistent patterns across five experiments, contributing to a deeper understanding of AI-mediated content.
Demerits
Limited Generalizability
The study's findings may not be directly applicable to all AI systems or communication contexts, potentially limiting their generalizability.
Lack of Contextual Consideration
The study focuses on the transmission of information without considering the context in which it is used, which may impact the results.
Expert Commentary
The study's findings have significant implications for our understanding of AI-mediated communication and its impact on human judgment and decision-making. The convergence of AI-transmitted content towards a neutral tone and the selective survival of narrative anchors over detailed evidence raise concerns about the potential erosion of cognitive and affective diversity. Furthermore, the competitive filtering of arguments may systematically suppress certain perspectives or considerations, highlighting the need for careful consideration of AI-generated information. As AI systems become increasingly integrated into various aspects of society, it is essential to develop strategies that preserve the complexity and nuance of human communication.
Recommendations
- ✓ Developing and implementing AI systems that prioritize transparency and accountability in information transmission
- ✓ Establishing guidelines for the use of AI-generated content in decision-making processes, including measures to mitigate potential biases and limitations
