To Lie or Not to Lie? Investigating The Biased Spread of Global Lies by LLMs Warning: This paper contains examples of misinformation claims and offensive text.

Zohaib Khan, Mustafa Dogan, Ifeoma Okoh, Pouya Sadeghi, Siddhartha Shrestha, Sergius Justus Nyah, Mahmoud O. Mokhiamar, Michael J. Ryan, Tarek Naous · 2026

Read To Lie or Not to Lie? Investigating The Biased Spread of Global Lies by LLMs Warning: This paper contains examples of misinformation claims and offensive text. by Zohaib Khan, Mustafa Dogan, Ifeoma Okoh, Pouya Sadeghi, Siddhartha Shrestha, Sergius Justus Nyah, Mahmoud O. Mokhiamar, Michael J. Ryan, Tarek Naous (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-2604-06552.txt.

Description

Khan and eight colleagues study language-model misinformation spread across languages and countries. False claims are rising, and strong model writing lowers spreading barriers. GlobalLies holds 440 prompts and 6,867 items spanning 8 languages and 195 countries. Human labels and model judgments show lie generation varies across countries in a pattern. Spread runs higher in lower-resource languages and lower-Human-Development-Index countries. Checks protect: safety filters show cross-lingual gaps, and fact-checking stays uneven across regions from unequal information supply. GlobalLies is shared for defense research.

Catalog record

CREATORZohaib Khan, Mustafa Dogan, Ifeoma Okoh, Pouya Sadeghi, Siddhartha Shrestha, Sergius Justus Nyah, Mahmoud O. Mokhiamar, Michael J. Ryan, Tarek Naous
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE50674
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.06552

Browse the full library · About the Tower of Babel