The Hidden Puppet Master: Predicting Human Belief Change in Manipulative LLM Dialogues
cs.CL
Submitted: 2026-03-21
Updated: 2026-10-12
Code: https://github.com/mitmedialab/llm-manipulation
Terminology
Sources
- AI Safety in Generative AI Large Language Models: A Survey
- ChatbotManip: A Dataset to Facilitate Evaluation and Oversight of Manipulative Chatbot Behaviour
- Large language models can effectively convince people to believe conspiracies
- The Fake Friend Dilemma: Relational Trust and the Political Economy of Conversational AI
- ManiTweet: A New Benchmark for Identifying Manipulation of News on Social Media
- A Meta-Analysis of the Persuasive Power of Large Language Models
- An Unsupervised Domain-Independent Framework for Automated Detection of Persuasion Tactics in Text
- How Do LLMs Persuade? Linear Probes Can Uncover Persuasion Dynamics in Multi-Turn Conversations
- SELF-PERCEPT: Introspection Improves Large Language Models' Detection of Multi-Person Mental Manipulation in Conversations
- Personalisation within bounds: A risk taxonomy and policy framework for the alignment of large language models with personalised feedback
- The History and Risks of Reinforcement Learning and Human Feedback
- LLM Can be a Dangerous Persuader: Empirical Study of Persuasion Safety in Large Language Models
- Human Decision-making is Susceptible to AI-driven Manipulation
- AI-LieDar: Examine the Trade-off Between Utility and Truthfulness in LLM Agents
- The Dark Side of AI Companionship: A Taxonomy of Harmful Algorithmic Behaviors in Human-AI Relationships
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering