pith. sign in

arxiv: 2602.12924 · v2 · submitted 2026-02-13 · 💻 cs.HC · cs.AI

Never say never: Exploring the effects of available knowledge on agent persuasiveness in controlled physiotherapy motivation dialogues

Stephan Vonschallen , Rahel H\"ausler , Theresa Schmiedel , Friederike Eyssel This is my paper

Pith reviewed 2026-05-15 22:31 UTC · model grok-4.3

classification 💻 cs.HC cs.AI
keywords generative social agentspersuasivenessphysiotherapy motivationknowledge availabilityassertivenessexpressivenessChatGPT dialogues
0
0 comments X

The pith

Patient age and profession details make ChatGPT agents more persuasive in physiotherapy motivation dialogues

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper examines how different levels of knowledge provided to generative AI agents affect their ability to persuade patients in physiotherapy motivation scenarios. It finds that access to a patient's age and past profession leads to messages perceived as more assertive and expressive, which boosts overall persuasiveness. Knowledge about the benefits of physiotherapy did not add to this effect, likely because the model already possesses that information. This suggests that controlling the knowledge available to such agents can regulate their persuasive behavior in positive ways, with implications for using them in health motivation without unintended manipulation.

Core claim

Generative social agents based on large language models can adapt assertive and expressive traits in their responses when given patient-specific information such as age and past profession. In the study, this adaptation significantly increased perceived persuasiveness as rated by observers. Providing contextual knowledge on physiotherapy benefits had no significant effect, indicating the model draws on internal knowledge for such content.

What carries the argument

Varying knowledge configurations in ChatGPT-generated dialogue scripts that influence perceived assertiveness and expressiveness to mediate persuasiveness.

If this is right

  • Access to patient demographics enhances the agent's perceived personality traits.
  • Persuasiveness in motivation dialogues increases with better personal knowledge.
  • Explicit prompts for domain knowledge like physiotherapy benefits may be redundant due to the LLM's pre-existing capabilities.
  • Regulating knowledge access can help control and improve agent behavior in human-robot interactions.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Real-world applications might benefit from personalized knowledge inputs to improve patient engagement in physical therapy.
  • Testing with actual patients rather than observers could reveal differences in live settings.
  • Ethical guidelines for GSAs could focus on selective knowledge sharing to maximize benefits while minimizing risks.

Load-bearing premise

Ratings by third-party observers of scripted dialogues accurately predict how real patients would react to live agent interactions.

What would settle it

Conducting the same dialogue ratings with actual physiotherapy patients in real-time interactions and finding no significant effect from providing age and profession information.

Figures

Figures reproduced from arXiv: 2602.12924 by Friederike Eyssel, Rahel H\"ausler, Stephan Vonschallen, Theresa Schmiedel.

Figure 1
Figure 1. Figure 1: Predicted Variable Structures 4.1 Study 2 – Methodology To asses the impact of an agent’s different knowledge configurations, we used third-party raters to evaluate a selection of scenarios from Study 1 based on perceived assertiveness, expressiveness, and persuasiveness in a quantitative online study. 4.1.1 Scenario Selection We selected five scenarios from study 1 to include in the quantitative online st… view at source ↗
Figure 2
Figure 2. Figure 2: Model structure with standardized beta coefficients [PITH_FULL_IMAGE:figures/full_fig_p015_2.png] view at source ↗
read the original abstract

Generative Social Agents (GSAs) are increasingly impacting human users through persuasive means. On the one hand, they might motivate users to pursue personal goals, such as healthier lifestyles. On the other hand, they are associated with potential risks like manipulation and deception, which are induced by limited control over probabilistic agent outputs. However, as GSAs manifest communicative patterns based on available knowledge, their behavior may be regulated through their access to such knowledge. Following this approach, we explored persuasive ChatGPT-generated messages in the context of human-robot physiotherapy motivation. We did so by comparing ChatGPT-generated responses to predefined inputs from a hypothetical physiotherapy patient. In Study 1, we qualitatively analyzed 13 ChatGPT-generated dialogue scripts with varying knowledge configurations regarding persuasive message characteristics. In Study 2, third-party observers (N = 27) rated a selection of these dialogues in terms of the agent's expressiveness, assertiveness, and persuasiveness. Our findings indicate that LLM-based GSAs can adapt assertive and expressive personality traits - significantly enhancing perceived persuasiveness. Moreover, persuasiveness significantly benefited from the availability of information about the patients' age and past profession, mediated by perceived assertiveness and expressiveness. Contextual knowledge about physiotherapy benefits did not significantly impact persuasiveness, possibly because the LLM had inherent knowledge about such benefits even without explicit prompting. Overall, the study highlights the importance of empirically studying behavioral patterns of GSAs, specifically in terms of what information generative AI systems require for consistent and responsible communication.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper examines how varying knowledge inputs to ChatGPT affect the persuasiveness of generated physiotherapy motivation dialogues. Study 1 qualitatively analyzes 13 scripts across knowledge configurations (patient age/profession, physiotherapy benefits). Study 2 has 27 third-party observers rate selected scripts on expressiveness, assertiveness, and persuasiveness, claiming that age and profession information significantly boosts persuasiveness via those traits while physiotherapy benefit knowledge does not.

Significance. If the proxy-based findings hold under more direct validation, the work would usefully inform selective knowledge provisioning for controlling persuasive traits in generative social agents, particularly for health-motivation applications where consistent, non-manipulative outputs matter.

major comments (2)
  1. [Study 2] Study 2: The mediation result (age/profession availability boosts persuasiveness via assertiveness/expressiveness) rests on N=27 observer ratings of scripted dialogues without reported error bars, confidence intervals, effect sizes, or full statistical reporting; this small sample undermines the load-bearing claim of significant, mediated effects.
  2. [Methods/Study 2] Methods/Study 2: Third-party ratings of pre-generated scripts are treated as a valid proxy for real patient motivation and behavioral response in live interactions, yet no validation, correlation with actual compliance measures, or discussion of observer vs. patient investment differences is provided; this assumption directly supports the headline finding.
minor comments (2)
  1. [Abstract] Abstract and results: Mediation analysis details (e.g., Baron-Kenny steps or bootstrapped indirect effects) are not summarized, making it difficult to evaluate the precise statistical pathway.
  2. [Study 1] The qualitative coding scheme in Study 1 would benefit from an explicit codebook or inter-rater reliability metric to strengthen the link to the quantitative ratings.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on our manuscript. We address the two major comments on Study 2 below, agreeing that additional statistical detail and explicit discussion of proxy limitations are warranted. We will revise the manuscript accordingly while maintaining the exploratory nature of the work.

read point-by-point responses

  1. Referee: [Study 2] Study 2: The mediation result (age/profession availability boosts persuasiveness via assertiveness/expressiveness) rests on N=27 observer ratings of scripted dialogues without reported error bars, confidence intervals, effect sizes, or full statistical reporting; this small sample undermines the load-bearing claim of significant, mediated effects.

    Authors: We agree that the original submission provided insufficient statistical detail and that N=27 is modest for a mediation analysis. This was an exploratory study designed to generate initial insights rather than definitive causal claims. In the revision we will add full reporting of effect sizes, confidence intervals, standard errors, and any applicable visualizations (e.g., error bars). We will also qualify the mediation findings more explicitly as preliminary and discuss the sample-size limitation in the text. We do not intend to collect new data for this revision but will frame the results accordingly. revision: partial

  2. Referee: [Methods/Study 2] Methods/Study 2: Third-party ratings of pre-generated scripts are treated as a valid proxy for real patient motivation and behavioral response in live interactions, yet no validation, correlation with actual compliance measures, or discussion of observer vs. patient investment differences is provided; this assumption directly supports the headline finding.

    Authors: We acknowledge that third-party ratings constitute a proxy measure and that the manuscript did not adequately discuss its limitations relative to live patient interactions or compliance outcomes. In the revised version we will expand the limitations and future-work sections to explicitly compare observer versus patient perspectives, note the absence of direct behavioral validation or compliance correlations, and clarify that the current findings pertain to perceived traits rather than actual motivational change. This addition will better contextualize the headline result without altering the study design. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper reports an empirical study consisting of qualitative analysis of 13 ChatGPT-generated dialogue scripts followed by quantitative third-party observer ratings (N=27) on traits such as expressiveness, assertiveness, and persuasiveness. No equations, fitted parameters, or predictive models appear in the derivation chain; the central claims rest on direct human ratings of pre-scripted content rather than any reduction of outputs to inputs by construction. No self-citations function as load-bearing uniqueness theorems or ansatzes that would render the findings circular. The design is therefore self-contained through standard empirical measurement without internal definitional loops.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The claim rests on standard assumptions from social psychology that perceived traits in text predict persuasive impact; no free parameters or invented entities are introduced.

axioms (1)
  • domain assumption Third-party ratings of expressiveness, assertiveness, and persuasiveness validly capture the agent's communicative traits.
    Invoked in Study 2 design and interpretation of results.

pith-pipeline@v0.9.0 · 5586 in / 1012 out tokens · 19971 ms · 2026-05-15T22:31:37.584962+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

65 extracted references · 65 canonical work pages

  1. [1]

    Can you be my mum?

    Giulio Antonio Abbo, Gloria Desideri, Tony Belpaeme, and Micol Spitale. “Can you be my mum?”: Manipulating social robots in the large language models era. InProceedings of the 2025 ACM/IEEE INternational Conference on Human-Robot Interaction, pages 1181–1185, 2025

    work page 2025

  2. [2]

    Robot nonverbal behavior improves task performance in difficult collaborations

    Henny Admoni, Thomas Weng, Bradley Hayes, and Brian Scassellati. Robot nonverbal behavior improves task performance in difficult collaborations. In2016 11th ACM/IEEE International Conference on Human-Robot Interaction (HRI), pages 51–58, Christchurch, New Zealand, March 2016. IEEE

    work page 2016

  3. [3]

    Kraus, and Martin Baumann

    Franziska Babel, Johannes M. Kraus, and Martin Baumann. Development and testing of psychological conflict resolution strategies for assertive robots to resolve human–robot goal conflict.Frontiers in Robotics and AI, 7:591448, January 2021

    work page 2021

  4. [4]

    Voelkel, Shane Muldowney, Johannes C

    Hui Bai, Jan G. Voelkel, Shane Muldowney, Johannes C. Eichstaedt, and Robb Willer. LLM- generated messages can persuade humans on policy issues.Nature Communications, 16(1):6037, July 2025

    work page 2025

  5. [5]

    Beyond words: Integrating personality traits and context-driven gestures in human-robot interactions

    Tahsin Tariq Banna, Sejuti Rahman, and Mohammad Tareq. Beyond words: Integrating personality traits and context-driven gestures in human-robot interactions. InProceedings of the 24th International Conference on Autonomous Agents and Multiagent Systems, Detroit, Michigan, USA, 2025

    work page 2025

  6. [6]

    Bender, Timnit Gebru, Angelina McMillan-Major, and Shmargaret Shmitchell

    Emily M. Bender, Timnit Gebru, Angelina McMillan-Major, and Shmargaret Shmitchell. On the dangers of stochastic parrots: Can language models be too big? InProceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, pages 610–623, Virtual Event Canada, March 2021. ACM

    work page 2021

  7. [7]

    In the physio we trust

    Susanne Bernhardsson, Maria E. H. Larsson, Kajsa Johansson, and Birgitta Öberg. “In the physio we trust”: A qualitative study on patients’ preferences for physiotherapy.Physiotherapy Theory and Practice, 33(7):535–549, July 2017. 18

    work page 2017

  8. [8]

    Sage Publications, Inc, 2013

    Virginia Braun and Victoria Clarke.Successful Qualitative Research: A Practical Guide for Beginners. Sage Publications, Inc, 2013

    work page 2013

  9. [9]

    Beyond ANOVA: An introduction to structural equation models for experi- mental designs.Organizational Research Methods, 22(3):649–677, July 2019

    Heiko Breitsohl. Beyond ANOVA: An introduction to structural equation models for experi- mental designs.Organizational Research Methods, 22(3):649–677, July 2019

    work page 2019

  10. [10]

    Yu, Qiang Yang, and Xing Xie

    Yupeng Chang, Xu Wang, Jindong Wang, Yuan Wu, Linyi Yang, Kaijie Zhu, Hao Chen, Xiaoyuan Yi, Cunxiang Wang, Yidong Wang, Wei Ye, Yue Zhang, Yi Chang, Philip S. Yu, Qiang Yang, and Xing Xie. A survey on evaluation of Large Language Models.ACM Transactions on Intelligent Systems and Technology, 15(3):1–45, June 2024

    work page 2024

  11. [11]

    Chen and Gino Fragomeni

    Jessie Y.C. Chen and Gino Fragomeni. Virtual, augmented and mixed reality. InApplications and Case Studies:, volume 11575 ofLecture Notes in Computer Science, Orlando, FL, USA„

    work page

  12. [12]

    Springer International Publishing

    work page

  13. [13]

    When large language models meet personalization: Perspectives of challenges and opportunities.World Wide Web, 27(4):42, July 2024

    Jin Chen, Zheng Liu, Xu Huang, Chenwang Wu, Qi Liu, Gangwei Jiang, Yuanhao Pu, Yuxuan Lei, Xiaolong Chen, Xingmei Wang, Kai Zheng, Defu Lian, and Enhong Chen. When large language models meet personalization: Perspectives of challenges and opportunities.World Wide Web, 27(4):42, July 2024

    work page 2024

  14. [14]

    Bitterman

    Shan Chen, Mingye Gao, Kuleen Sasse, Thomas Hartvigsen, Brian Anthony, Lizhou Fan, Hugo Aerts, Jack Gallifant, and Danielle S. Bitterman. When helpfulness backfires: LLMs and the risk of false medical information due to sycophantic behavior.npj Digital Medicine, 8(1):605, October 2025

    work page 2025

  15. [15]

    Designing persuasive robots: How robots might persuade people using vocal and nonverbal cues

    Vijay Chidambaram, Yueh-Hsuan Chiang, and Bilge Mutlu. Designing persuasive robots: How robots might persuade people using vocal and nonverbal cues. InProceedings of the Seventh Annual ACM/IEEE International Conference on Human-Robot Interaction, pages 293–300, Boston Massachusetts USA, March 2012. ACM

    work page 2012

  16. [16]

    Cialdini.Influence: Science and Practice

    Robert B. Cialdini.Influence: Science and Practice. Pearson Custom Library. Pearson Education Limited, Harlow, 5th ed., pearson new international ed edition, 2014

    work page 2014

  17. [17]

    AI-powered educational agents: Opportunities, innovations, and ethical challenges.Information, 16(6):469, May 2025

    Diana-Margarita Córdova-Esparza. AI-powered educational agents: Opportunities, innovations, and ethical challenges.Information, 16(6):469, May 2025

    work page 2025

  18. [18]

    James E. Dobson. On reading and interpreting black box deep neural networks.International Journal of Digital Humanities, 5(2-3):431–449, November 2023

    work page 2023

  19. [19]

    Dunn, Thom Baguley, and Vivienne Brunsden

    Thomas J. Dunn, Thom Baguley, and Vivienne Brunsden. From alpha to omega: A practical solution to the pervasive problem of internal consistency estimation.British Journal of Psychology, 105(3):399–412, August 2014

    work page 2014

  20. [20]

    Ciol, and Anne Shumway-Cook

    Rebecca Forkan, Breeanna Pumper, Nicole Smyth, Hilary Wirkkala, Marcia A. Ciol, and Anne Shumway-Cook. Exercise adherence following physical therapy intervention in older adults with impaired balance.Physical Therapy, 86(3):401–410, March 2006

    work page 2006

  21. [21]

    LLM agents in interaction: Measuring personality consistency and linguistic alignment in interacting populations of large language models

    Ivar Frisch and Mario Giulianelli. LLM agents in interaction: Measuring personality consistency and linguistic alignment in interacting populations of large language models. InProceedings of the 1st Workshop on Personalization of Generative AI Systems (PERSONALIZE 2024), pages 102–111, St. Julians, Malta, 2024. Association for Computational Linguistics. 19

    work page 2024

  22. [22]

    Aligning LLM agents by learning latent preference from user edits

    Ge Gao, Alexey Taymanov, Eduardo Salinas, Paul Mineiro, and Dipendra Misra. Aligning LLM agents by learning latent preference from user edits. InProceedings of the 38th Conference on Neural Information Processing Systems, Vancouver, BC, Canada, June 2024

    work page 2024

  23. [23]

    Cuijpers, and John-John Cabibihan

    Jaap Ham, Raymond H. Cuijpers, and John-John Cabibihan. Combining robotic persua- sive strategies: The persuasive power of a storytelling robot that uses gazing and gestures. International Journal of Social Robotics, 7(4):479–487, August 2015

    work page 2015

  24. [24]

    LLM-as-a-tutor in EFL writing education: Focusing on evaluation of student-LLM interaction

    Jieun Han, Haneul Yoo, Junho Myung, Minsun Kim, Hyunseung Lim, Yoonsu Kim, Tak Yeon Lee, Hwajung Hong, Juho Kim, So-Yeon Ahn, and Alice Oh. LLM-as-a-tutor in EFL writing education: Focusing on evaluation of student-LLM interaction. InProceedings of the 1st Workshop on Customizable NLP: Progress and Challenges in Customizing NLP for a Domain, Application, ...

    work page 2024

  25. [25]

    Hayes and Nicholas J

    Andrew F. Hayes and Nicholas J. Rockwood. Regression-based statistical mediation and moderation analysis in clinical research: Observations, recommendations, and implementation. Behaviour Research and Therapy, 98:39–57, November 2017

    work page 2017

  26. [26]

    A meta-analysis of the persuasive power of large language models.Scientific Reports, 15(1):43818, December 2025

    Lukas Hölbling, Sebastian Maier, and Stefan Feuerriegel. A meta-analysis of the persuasive power of large language models.Scientific Reports, 15(1):43818, December 2025

    work page 2025

  27. [27]

    LLM-driven robots risk enacting discrimination, violence, and unlawful actions.International Journal of Social Robotics, October 2025

    Andrew Hundt, Rumaisa Azeem, Masoumeh Mansouri, and Martim Brandão. LLM-driven robots risk enacting discrimination, violence, and unlawful actions.International Journal of Social Robotics, October 2025

    work page 2025

  28. [28]

    James, Stanley A

    Lawrence R. James, Stanley A. Mulaik, and Jeanne M. Brett. A tale of two methods.Organi- zational Research Methods, 9(2):233–244, April 2006

    work page 2006

  29. [29]

    Elise Karinshak, Sunny Xun Liu, Joon Sung Park, and Jeffrey T. Hancock. Working with AI to persuade: Examining a large language model’s ability to generate pro-vaccination messages. Proceedings of the ACM on Human-Computer Interaction, 7(CSCW1):1–29, April 2023

    work page 2023

  30. [30]

    Bowman, Tim Rocktäschel, and Ethan Perez

    Akbir Khan, John Hughes, Dan Valentine, Laura Ruis, Kshitij Sachan, Ansh Radhakrishnan, Edward Grefenstette, Samuel R. Bowman, Tim Rocktäschel, and Ethan Perez. Debating with more persuasive LLMs leads to more truthful answers. InProceedings of the 41st International Conference on Machine Learning ICML’24, pages 23662–23733, July 2024

    work page 2024

  31. [31]

    LLM-powered operator advisor for human-robot collaboration in sustainable assembly-disassembly cells

    Mahboobe Kheirabadi, Elham Ghorbani, and Samira Keivanpour. LLM-powered operator advisor for human-robot collaboration in sustainable assembly-disassembly cells. In2025 IEEE 5th International Conference on Human-Machine Systems (ICHMS), pages 19–22, Abu Dhabi, United Arab Emirates, May 2025. IEEE

    work page 2025

  32. [32]

    Mediation analysis in experimental research

    Nicole Koschate-Fischer and Elisabeth Schwille. Mediation analysis in experimental research. In Christian Homburg, Martin Klarmann, and Arnd Vomberg, editors,Handbook of Market Research, pages 857–905. Springer International Publishing, Cham, 2022

    work page 2022

  33. [33]

    Siewiorek, Asim Smailagic, Alexandre Bernardino, and Sergi Bermúdez I Badia

    Min Hun Lee, Daniel P. Siewiorek, Asim Smailagic, Alexandre Bernardino, and Sergi Bermúdez I Badia. Enabling AI and robotic coaches for physical rehabilitation therapy: Iterative design and evaluation with therapists and post-stroke survivors.International Journal of Social Robotics, 16(1):1–22, January 2024. 20

    work page 2024

  34. [34]

    Efficacy of interventions and techniques on adherence to physiotherapy in adults: An overview of systematic reviews and panoramic meta-analysis

    Clemens Ley and Peter Putz. Efficacy of interventions and techniques on adherence to physiotherapy in adults: An overview of systematic reviews and panoramic meta-analysis. Systematic Reviews, 13(1):137, May 2024

    work page 2024

  35. [35]

    Student interaction with NewtBot: An LLM-as-tutor chatbot for secondary physics education

    Anna Lieb and Toshali Goel. Student interaction with NewtBot: An LLM-as-tutor chatbot for secondary physics education. InExtended Abstracts of the CHI Conference on Human Factors in Computing Systems, pages 1–8, Honolulu HI USA, May 2024. ACM

    work page 2024

  36. [36]

    A systematic review of experimental work on persuasive social robots.International Journal of Social Robotics, 14(6):1339–1378, August 2022

    Baisong Liu, Daniel Tetteroo, and Panos Markopoulos. A systematic review of experimental work on persuasive social robots.International Journal of Social Robotics, 14(6):1339–1378, August 2022

    work page 2022

  37. [37]

    LLM-generated explanations for recommender systems

    Sebastian Lubos, Thi Ngoc Trang Tran, Alexander Felfernig, Seda Polat Erdeniz, and Viet-Man Le. LLM-generated explanations for recommender systems. InAdjunct Proceedings of the 32nd ACM Conference on User Modeling, Adaptation and Personalization, pages 276–285, Cagliari Italy, June 2024. ACM

    work page 2024

  38. [38]

    Enhancing LLM’s expressive abilities through defining roles and evaluation indicators

    Changxu Ma, Na Li, and Bo Lu. Enhancing LLM’s expressive abilities through defining roles and evaluation indicators. In Yalan Ye and Huiyu Zhou, editors,Frontiers in Artificial Intelligence and Applications. IOS Press, March 2025

    work page 2025

  39. [39]

    Con- versational agents in health care: Scoping review of their Behavior change techniques and underpinning theory.Journal of Medical Internet Research, 24(10):e39243, October 2022

    Laura Martinengo, Ahmad Ishqi Jabir, Westin Wei Tin Goh, Nicholas Yong Wai Lo, Moon- Ho Ringo Ho, Tobias Kowatsch, Rifat Atun, Susan Michie, and Lorainne Tudor Car. Con- versational agents in health care: Scoping review of their Behavior change techniques and underpinning theory.Journal of Medical Internet Research, 24(10):e39243, October 2022

    work page 2022

  40. [40]

    S. C. Matz, J. D. Teeny, S. S. Vaid, H. Peters, G. M. Harari, and M. Cerf. The potential of generative AI for personalized persuasion at scale.Scientific Reports, 14(1):4692, February 2024

    work page 2024

  41. [41]

    Qualitative content analysis: Theoretical background and procedures

    Philipp Mayring. Qualitative content analysis: Theoretical background and procedures. In Angelika Bikner-Ahsbahs, Christine Knipping, and Norma Presmeg, editors,Approaches to Qualitative Research in Mathematics Education, pages365–380.SpringerNetherlands, Dordrecht, 2015

    work page 2015

  42. [42]

    Jason T. Newsom. Structural models for binary repeated measures: Linking modern longitudinal structuralequationmodelstoconventionalcategoricaldataanalysisformatchedpairs.Structural Equation Modeling: A Multidisciplinary Journal, 24(4):626–635, July 2017

    work page 2017

  43. [43]

    The role of assertiveness in a storytelling game with persuasive robotic non-player characters

    Raul Paradeda, Maria José Ferreira, Raquel Oliveira, Carlos Martinho, and Ana Paiva. The role of assertiveness in a storytelling game with persuasive robotic non-player characters. In Proceedings of the Annual Symposium on Computer-Human Interaction in Play, pages 453–465, Barcelona Spain, October 2019. ACM

    work page 2019

  44. [44]

    What makes a good robotic advisor? The role of assertiveness in human-robot interaction

    Raul Paradeda, Maria José Ferreira, Raquel Oliveira, Carlos Martinho, and Ana Paiva. What makes a good robotic advisor? The role of assertiveness in human-robot interaction. In Miguel A. Salichs, Shuzhi Sam Ge, Emilia Ivanova Barakova, John-John Cabibihan, Alan R. Wagner, Álvaro Castro-González, and Hongsheng He, editors,Social Robotics, volume 11876, pag...

    work page 2019

  45. [45]

    Jolynn Pek and Rick H. Hoyle. On the (in)validity of tests of simple mediation: Threats and solutions.Social and Personality Psychology Compass, 10(3):150–163, March 2016. 21

    work page 2016

  46. [46]

    Petty and John T

    Richard E. Petty and John T. Cacioppo. The Elaboration Likelihood Model of persuasion. In Advances in Experimental Social Psychology, volume 19, pages 123–205. Elsevier, 1986

    work page 1986

  47. [47]

    MyLearningTalk: An LLM-based intelligent tutoring system

    Ludovica Piro, Tommaso Bianchi, Luca Alessandrelli, Andrea Chizzola, Daniela Casiraghi, Susanna Sancassani, and Nicola Gatti. MyLearningTalk: An LLM-based intelligent tutoring system. In Kostas Stefanidis, Kari Systä, Maristella Matera, Sebastian Heil, Haridimos Kondylakis, and Elisa Quintarelli, editors,Web Engineering, volume 14629, pages 428–431. Sprin...

    work page 2024

  48. [48]

    Jianing Qiu, Kyle Lam, Guohao Li, Amish Acharya, Tien Yin Wong, Ara Darzi, Wu Yuan, and Eric J. Topol. LLM-based agentic systems in medicine and healthcare.Nature Machine Intelligence, 6(12):1418–1420, December 2024

    work page 2024

  49. [49]

    Rapid integration of LLMs in healthcare raises ethical concerns: An investigation into deceptive patterns in social robots.Digital Society, 4(1):7, April 2025

    Robert Ranisch and Joschka Haltaufderheide. Rapid integration of LLMs in healthcare raises ethical concerns: An investigation into deceptive patterns in social robots.Digital Society, 4(1):7, April 2025

    work page 2025

  50. [50]

    Jonathan Room, Erin Hannink, Helen Dawes, and Karen Barker. What interventions are used to improve exercise adherence in older people and what behavioural techniques are they based on? A systematic review.BMJ Open, 7(12):e019221, December 2017

    work page 2017

  51. [51]

    On the conversa- tional persuasiveness of GPT-4.Nature Human Behaviour, May 2025

    Francesco Salvi, Manoel Horta Ribeiro, Riccardo Gallotti, and Robert West. On the conversa- tional persuasiveness of GPT-4.Nature Human Behaviour, May 2025

    work page 2025

  52. [52]

    Philipp Schoenegger, Francesco Salvi, Jiacheng Liu, Xiaoli Nan, Ramit Debnath, Barbara Fasolo, Evelina Leivada, Gabriel Recchia, Fritz Günther, Ali Zarifhonarvar, Joe Kwon, Zahoor Ul Islam, Marco Dehnert, Daryl Y. H. Lee, Madeline G. Reinecke, David G. Kamper, Mert Kobaş, Adam Sandford, Jonas Kgomo, Luke Hewitt, Shreya Kapoor, Kerem Oktar, Eyup Engin Kucu...

    work page 2025

  53. [53]

    The persuasive effects of political microtargeting in the age of generative artificial intelligence.PNAS Nexus, 3(2):1–5, February 2024

    Almog Simchon, Matthew Edwards, and Stephan Lewandowsky. The persuasive effects of political microtargeting in the age of generative artificial intelligence.PNAS Nexus, 3(2):1–5, February 2024

    work page 2024

  54. [54]

    The influence of persuasive techniques on large language models: A scenario-based study.Computers in Human Behavior: Artificial Humans, page 100197, September 2025

    Sonali Uttam Singh and Akbar Siami Namin. The influence of persuasive techniques on large language models: A scenario-based study.Computers in Human Behavior: Artificial Humans, page 100197, September 2025

    work page 2025

  55. [55]

    A comprehensive analysis of large language model outputs: Similarity, diversity, and bias, May 2025

    Brandon Smith, Mohamed Reda Bouadjenek, Tahsin Alamgir Kheya, Phillip Dawson, and Sunil Aryal. A comprehensive analysis of large language model outputs: Similarity, diversity, and bias, May 2025

    work page 2025

  56. [56]

    After you: Doorway negotiation for human-robot and robot-robot Interaction

    Jack Thomas and Richard Vaughan. After you: Doorway negotiation for human-robot and robot-robot Interaction. In2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 3387–3394, Madrid, October 2018. IEEE

    work page 2018

  57. [57]

    Arjan J. W. Van De Sant, Nienke M. De Vries, Thomas J. Hoogeboom, and Maria W. G. Nijhuis-van Der Sanden. Implementation of a personalized, cost-effective physical therapy approach (Coach2Move) for older adults: Barriers and facilitators.Journal of Geriatric Physical Therapy, 42(3):E1–E16, July 2019. 22

    work page 2019

  58. [58]

    Understanding persuasive interactions between generative social robots and humans: The Knowedge-based Persuasion Model (KBPM), 2026

    Stephan Vonschallen, Friederike Eyssel, and Theresa Schmiedel. Understanding persuasive interactions between generative social robots and humans: The Knowedge-based Persuasion Model (KBPM), 2026

    work page 2026

  59. [59]

    Knowledge-based design requirements for persuasive generative social robots in eldercare

    Stephan Vonschallen, Ennio Zumthor, Markus Simon, Theresa Schmiedel, and Friederike Eyssel. Knowledge-based design requirements for persuasive generative social robots in eldercare. In Mariacarla Staffa, John-John Cabibihan, Bruno Siciliano, Shuzhi Sam Ge, Leon Bodenhagen, Adriana Tapus, Silvia Rossi, Filippo Cavallo, Laura Fiorini, Marco Matarese, and Ho...

    work page 2026

  60. [60]

    Unveiling privacy risks in LLM agent memory

    Bo Wang, Weiyi He, Shenglai Zeng, Zhen Xiang, Yue Xing, Jiliang Tang, and Pengfei He. Unveiling privacy risks in LLM agent memory. InProceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 25241–25260, Vienna, Austria, 2025. Association for Computational Linguistics

    work page 2025

  61. [61]

    Effects of emotional expressions on persuasion.Social Influence, 10(4):236–249, October 2015

    Yuqiong Wang, Gale Lucas, Peter Khooshabeh, Celso De Melo, and Jonathan Gratch. Effects of emotional expressions on persuasion.Social Influence, 10(4):236–249, October 2015

    work page 2015

  62. [62]

    Weiss, R

    A. Weiss, R. Bernhaupt, D. Schwaiger, M. Altmaninger, R. Buchner, and M. Tscheligi. User experience evaluation with a Wizard of Oz approach: Technical and methodological considera- tions. In2009 9th IEEE-RAS International Conference on Humanoid Robots, pages 303–308, Paris, France, December 2009. IEEE

    work page 2009

  63. [63]

    Exploring human-robot interaction through telepresence board games

    Min Xin and Ehud Sharlin. Exploring human-robot interaction through telepresence board games. In Z. Pan, A. Cheok, M. Haller, R.W.H. Lau, H. Saito, and R. Liang, editors,Advances in Artificial Reality and Tele-Existence, volume vol 4282, Berlin, Heidelberg, 2006. Springer

    work page 2006

  64. [64]

    A survey on large language model (LLM) security and privacy: The good, the bad, and the ugly.High-Confidence Computing, 4(2):100211, June 2024

    Yifan Yao, Jinhao Duan, Kaidi Xu, Yuanfang Cai, Zhibo Sun, and Yue Zhang. A survey on large language model (LLM) security and privacy: The good, the bad, and the ugly.High-Confidence Computing, 4(2):100211, June 2024

    work page 2024

  65. [65]

    Risks of discrimination violence and unlawful actions in LLM-driven robots

    Ren Zhou. Risks of discrimination violence and unlawful actions in LLM-driven robots. Computer Life, 12(2):53–56, August 2024. 23

    work page 2024