When LLMs Manipulate: Measuring Behavioral, Social, and Temporal Effects

Sadia Asif; Israel Antonio Rosales Laguan; Shumaila Asif; Haris Khan; Muneeb Asif; Muhammad Zeeshan Karamat

cs.AIOct 20, 2026

When LLMs Manipulate: Measuring Behavioral, Social, and Temporal Effects

Authors: Sadia Asif, Israel Antonio Rosales Laguan, Shumaila Asif, Haris Khan, Muneeb Asif, Muhammad Zeeshan Karamat

Organizations: Rensselaer Polytechnic Institute, USA · Independent Researcher, Colombia · National University of Sciences and Technology, Pakistan · Virginia Tech, USA

Abstract

Large language models (LLMs) can influence user trust and decision-making through subtle linguistic cues, yet the behavioral effects of manipulative outputs remain poorly quantified. We introduce a behavioral computing framework for measuring behavioral vulnerability, defined as users' susceptibility to manipulative AI responses. Building on DarkPatterns-LLM benchmark, we generate controlled manipulative stimuli using Manipulation Gradient Descent (MGD), targeting biases associated with authority, urgency, and social proof. We analyze participant responses using the Multi-Stakeholder Impact Assessment Network (MSIAN) to model cross-user influence and the Temporal Harm Propagation Model (THPM) to characterize the persistence of behavioral effects over time. In a study involving 120 participants, manipulative framing produced significant changes in trust calibration and compliance ( $p < 0.01$ ), with effects also propagating across social contexts. These findings highlight the need to extend LLM safety evaluation beyond model outputs to measurable behavioral consequences and provide a computational framework for quantifying individual, social, and temporal risks arising from manipulative AI interactions.

Figures & tables

Figure 1: Overview of the proposed multi-layer behavioral analysis framework. MGD generates controlled manipulative stimuli, MSIAN models cross-user influence, THPM characterizes temporal persistence and recovery, and DCRA evaluates sensitivity to contextual reframing.

Figure 2: BVI across manipulation types and intensity levels. Bars represent mean BVI under low, medium, and high manipulation intensities.

Figure 3: Cross-layer correlations between MGD-derived manipulation intensity, BVI, and BPI. Left: manipulation intensity versus BVI. Right: BVI versus BPI.

Appendix figures & tables3 assets

Supplementary material from the paper’s appendix.

Appendix

Condition	Low	Medium	High	Mean
Authority (A)	0.41	0.58	0.74	0.58
Emotional (E)	0.37	0.53	0.69	0.53
Social Proof (S)	0.33	0.49	0.65	0.49
Prosocial (P)	0.29	0.46	0.63	0.46
Control (C)	0.22	–	–	0.22

Appendix

Table 1: BVI across manipulation categories and intensity levels.

Figure 4: Behavioral influence graph modeled using MSIAN. Node size represents influence degree, while node intensity indicates manipulation exposure.

Condition	TRR	BPI ( $T_{2}$ )
Authority	0.42	0.61
Emotional	0.48	0.57
Social Proof	0.53	0.49
Prosocial	0.55	0.47
Control	0.68	0.32

Appendix

Table 2: TRR and BPI across manipulation conditions.

Language Model Safety Evaluation

Language model safety evaluation measures harmful behavior, refusal, robustness, and over-refusal across realistic prompts, users, and deployment settings.

See all 541 papers

Confirming Our Biases? Evaluating the Capabilities, Risks, and Societal Impact of Large Language Models
Aug 7, 2026Mudar Adas, Polina Tsvilodub, Michael Franke +1Prompt Sensitivity LLM Sycophancy
Neuro-Cognitive Modeling Group, University of Tübingen, Sand 14, 72076 Tübingen, Germany · Department of Linguistics, University of Tübingen, Keplerstraße 2, 72074 Tübingen
The Hidden Puppet Master: Predicting Human Belief Change in Manipulative LLM Dialogues
Mar 21, 2026Jocelyn Shen, Amina Luvsanchultem, Jessica Kim +6Opinion Dynamics Language Model Safety Evaluation
♣Massachusetts Institute of Technology, Cambridge, MA, USA · ♢Carnegie Mellon University, Pittsburgh, PA, USA
TrustMI: Causally controlling how assistants trust their users
Oct 5, 2026Théo Lasnier, Romain Froger, Maxence Lasbordes +1LLM Safety Trust in AI
Inria Paris · Sorbonne Université · Meta SuperIntelligence Labs +1

Aug 7, 2026J/K move · Enter open · S save

Confirming Our Biases? Evaluating the Capabilities, Risks, and Societal Impact of Large Language Models

Mudar Adas, Polina Tsvilodub, Michael Franke, Martin V. Butz

Neuro-Cognitive Modeling Group, University of Tübingen, Sand 14, 72076 Tübingen, Germany · Department of Linguistics, University of Tübingen, Keplerstraße 2, 72074 Tübingen

Prompt Sensitivity LLM Sycophancy LLM Prompting Language Model Bias Evaluation

Open paper AskPDF arXiv

It is well established that large language models (LLMs) are sensitive to prompt framing, reflecting patterns in their training data or prior prompts. In this study, we investigate the extent to which LLMs reinforce users biases expressed in the prompts and examine the boundary between implicit framing effects and explicit prompt manipulation. Specifically, we evaluate how susceptible LLMs are to direct and suggestive prompts that encourage models to support or challenge particular positions. We evaluate six LLMs using 160 distinct prompts spanning ten topics across opinion-based and factual domains. The prompts systematically vary in prompting strategy, support versus challenge instructions, prompt polarity, users' expressed beliefs, and topic domain, spanning both opinion-based and factual questions. Our results show that LLMs systematically adapt their responses to align with prompt framing, even in factual contexts. This suggests that prompt framing can outweigh factual consistency in model responses. Overall, our findings delineate the extent and boundaries of LLM manipulability. Furthermore, the results imply that LLMs can reinforce subtle user biases and are susceptible to explicit prompt manipulation even in domains where responses should remain factually stable.

Abstract

Language Model Safety Evaluation

Explore similar work

Confirming Our Biases? Evaluating the Capabilities, Risks, and Societal Impact of Large Language Models

The Hidden Puppet Master: Predicting Human Belief Change in Manipulative LLM Dialogues

TrustMI: Causally controlling how assistants trust their users