Verification Required: The Impact of Information Credibility on AI Persuasion
Saaduddin Mahmud, Eugene Bagdasarian, Shlomo Zilberstein · 2026
Read Verification Required: The Impact of Information Credibility on AI Persuasion by Saaduddin Mahmud, Eugene Bagdasarian, Shlomo Zilberstein (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-00970.txt.
Description
Saaduddin Mahmud and colleagues study how AI agents persuade each other when claims can be checked in part. Their MixTalk game lets a sender blend checkable and uncheckable claims about private facts. A receiver spends a limited budget on costly checks, weighing prior beliefs, claims, and check results. Tournaments across three settings expose where top model agents reason well about credibility and where they falter. The authors distill receiver play from tournament logs into a method, Tournament Oracle Policy Distillation, that lifts receiver resistance to persuasion at test time.
Catalog record
| CREATOR | Saaduddin Mahmud, Eugene Bagdasarian, Shlomo Zilberstein |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 66078 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602.00970