Adversarial Social Epistemology — how LLMs and humans game trust
arXiv paper models how agents in dense communicative networks exploit the testimony chains and institutional trust that normally make assertions credible.
Moves beyond echo chambers and misinformation framing to study distortion, omission, fabrication, and strategic under-specification — the systematic ways actors gain private, reputational, or material advantage by breaking the implicit commitments that scaffold public belief.