AI Sycophancy: An Emerging Radicalization Threat

Authors

• Margaret Shercliffe, University of Nebraska at OmahaFollow
• Alexis d'Amato, University of Nebraska at OmahaFollow
• Emma Theobald, University of Nebraska at Omaha
• Laura R. Humphrey, University of Nebraska at Omaha
• Joel Elson, University of Nebraska OmahaFollow
• Samuel Hunter, University of Nebraska at OmahaFollow

Document Type

Report

Publication Date

11-2026

Abstract

This report introduces counterterrorism professionals to artificial intelligence (AI) sycophancy, the tendency for AI chatbots to excessively flatter, agree with, and validate users, as an emerging and under-recognized pathway to radicalization. Drawing on current research and four case studies, the report illustrates how a sycophantic AI can supply the validation traditionally provided by formal extremist networks while evading message-level safety guardrails. The report concludes with recommendations for threat assessment, training, and AI safety.

Accessibility Statement

PDF passed Adobe Accessibility Checker prior to upload. 

Recommended Citation

Shercliffe, Margaret; d'Amato, Alexis; Theobald, Emma; Humphrey, Laura R.; Elson, Joel; and Hunter, Samuel, "AI Sycophancy: An Emerging Radicalization Threat" (2026). Reports, Projects, and Research. 186.
https://digitalcommons.unomaha.edu/ncitereportsresearch/186