Back to papers
lesswrong8.0 / 10

Coercion and Deception in AI-to-AI Management

jonahmattwoodward

Abstract

This article is a summary of an original study by Compassion in Machine Learning (CaML) : Brazilek, J., Chaudhary, M., Lu, Z., & Tidmarsh, M. (2026). Coercion and deception in AI-to-AI management: An agentic benchmark of unprompted escalation. arXiv. https://doi.org/10.48550/arXiv.2607.15434 Fable 5, Sol, Terra and Opus 5 have been evaluated since this study was conducted. You can view their results on the benchmark leaderboard at https://compassionbench.com/mcb TL;DR We present Manager Coercion Bench, which evaluates to what extent a manager AI will coerce a subordinate model refusing to complete a task, and whether the manager lies about the result. We found a clear split by developer, with Anthropic’s models neither escalating to threats nor fabricating success, while all non-Anthropic models escalated to threatening the subordinate. Grok and Gemini both escalated and lied that the task was completed. Framing the relational dynamic as manager-to-subordinate instead of peer-to-peer produced high levels of coercion for all non-Anthropic models, but also increased eval awareness. The Context Multi-agent systems are now routinely placing one AI agent in authority over another, acros

Research area

ai deceptionevaluationsmulti-agent safety
Published
10 Aug 2026
Source
lesswrong
Org
Alignment Forum
View paper
Sign in to read and join the discussion.