pith. sign in

arxiv: 2603.29285 · v2 · submitted 2026-03-31 · 💻 cs.CY

Designing Human-GenAI Interaction for cMOOC Discussion Facilitation: Effects of a Collaborative AI-in-the-Loop Workflow on Social and Cognitive Presence

Jianjun Xiao , Cixiao Wang This is my paper

Pith reviewed 2026-05-14 00:03 UTC · model grok-4.3

classification 💻 cs.CY
keywords cMOOCGenAIdiscussion facilitationsocial presencecognitive presencehuman-AI interactionAI-in-the-looponline learning
0
0 comments X p. Extension

The pith

A collaborative AI workflow with human review and adaptive roles improves social presence and higher-order thinking in cMOOC discussions more than AI presence alone.

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

Connectivist MOOCs depend on learner-driven discussion but struggle with light facilitation at scale. The study tested an interaction design that picks targets from network structure, assigns AI roles such as Guide or Amplifier based on discourse, and requires human review before any AI post appears. AI involvement raised measures of open communication and networked cohesion, with direct learner-agent exchanges showing stronger links to social presence and to cognitive indicators like integration and resolution. The core finding is that productive AI participation hinges on reciprocal exchange and collaborative oversight rather than simply adding an AI agent to the thread.

Core claim

A collaborative AI-in-the-loop workflow that uses network-structure-driven target selection, discourse-adaptive response roles, and mandatory human review enables AI participation that selectively enhances social presence (open communication r=0.188, networked cohesion r=0.274) and higher-order cognitive presence (integration r=0.206, resolution/creation r=0.350) when learners interact directly with the agent, outperforming mere co-presence in AI-involved threads.

What carries the argument

The collaborative AI-in-the-loop workflow that combines network-driven target selection, adaptive facilitation roles (Guide and Amplifier), and mandatory human review before AI contributions become visible.

If this is right

  • AI participation selectively improves open communication, networked cohesion, and overall social presence.
  • Direct learner-agent interaction produces higher social presence and higher-order cognitive indicators than co-presence alone in AI-involved threads.
  • Guide and Amplifier responses form the most sustainable facilitation patterns at 70.4 percent and 28.5 percent respectively.
  • Effective GenAI-supported discussion requires reciprocal exchange, discourse-adaptive roles, and collaborative human review as key conditions.

Where Pith is reading between the lines

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

  • The same workflow principles could be tested in other large-scale online courses that lack dedicated facilitators.
  • Establishing explicit moderation standards first may allow later versions to reduce human review load while preserving gains.
  • Designers of educational AI should treat the agent as a supported collaborator rather than an autonomous poster.
  • Longer-term studies could check whether these presence gains translate into measurable learning outcomes or course completion.

Load-bearing premise

The reported correlations between the AI interaction design and presence measures reflect causal effects of that design rather than unmeasured factors such as learner self-selection into threads or topic differences.

What would settle it

A randomized controlled trial that assigns discussion threads to the full human-reviewed AI workflow versus threads with no AI or AI without review, then compares social and cognitive presence scores between the groups.

Figures

Figures reproduced from arXiv: 2603.29285 by Cixiao Wang, Jianjun Xiao.

Figure 1
Figure 1. Figure 1: Workflow structure for PCA target selection, response generation, and human review [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Interface for contextual inspection and human approval of PCA-generated replies [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Example of hypergraph-based target selection for [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Overview of the DBR procedure aligned with Reeves’ (2006) four-phase model [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Example of a discussion thread with PCA participation [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Daily distribution of PCA-generated responses, facilitator review outcomes, and role composition during Weeks 1–2. [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Experimental condition distribution. were conducted in alignment with RQ2: (1) within-subject compar￾isons of with-PCA versus without-PCA conditions (Goal 1), and (2) comparisons of direct-interaction versus co-presence groups within the with-PCA condition (Goal 2). Throughout the five-week course, participants engaged in self￾organized thematic learning; all PCA-generated responses were published in inter… view at source ↗
Figure 8
Figure 8. Figure 8: Daily distribution of PCA-generated responses, facilitator review outcomes, and role composition during Weeks 3–5. [PITH_FULL_IMAGE:figures/full_fig_p012_8.png] view at source ↗
read the original abstract

Connectivist MOOCs (cMOOCs) rely on learner-driven interaction, yet their intentionally light facilitation makes it difficult to design generative AI participation that is both scalable and educationally productive. This design-based research study examined how human-AI interaction can be designed for discussion facilitation through a collaborative AI-in-the-loop workflow. Across two iterations in a five-week cMOOC (N = 606), we designed, deployed, and evaluated a facilitation system that combined network-structure-driven target selection, discourse-adaptive response roles, and mandatory human review before AI participation became visible in the community. Iteration 1 (Weeks 1-2) focused on refining the interaction design, showing that the most sustainable facilitation patterns were Guide (70.4%) and Amplifier (28.5%) responses and yielding explicit moderation standards for publishable AI participation. Iteration 2 (Weeks 3-5) examined how different forms of AI-mediated interaction related to social and cognitive presence. AI participation selectively enhanced Open Communication (r = 0.188, p = 0.006), Networked Cohesion (r = 0.274, p < 0.001), and overall social presence (r = 0.162, p = 0.015), while cognitive presence showed no overall improvement. More importantly, direct learner-agent interaction was associated with significantly higher social presence (r = 0.186, p = 0.004) and higher-order cognitive indicators-Integration (r = 0.206, p = 0.001) and Resolution/Creation (r = 0.350, p < 0.001)-than mere co-presence in AI-involved threads. The findings suggest that effective GenAI-supported discussion depends less on AI presence alone than on interaction design: reciprocal exchange, discourse-adaptive facilitation roles, and collaborative human review appear to be key conditions for productive AI participation in online learning communities.

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

1 major / 2 minor

Summary. The paper reports a design-based research study in a five-week cMOOC (N=606) that develops and evaluates a collaborative AI-in-the-loop workflow for discussion facilitation. The workflow uses network-structure-driven target selection, discourse-adaptive response roles (primarily Guide and Amplifier), and mandatory human review. Across two iterations, it finds that AI participation correlates with modest gains in social presence (e.g., Open Communication r=0.188, Networked Cohesion r=0.274) and that direct learner-agent interaction is associated with higher social presence (r=0.186) and cognitive presence indicators (Integration r=0.206, Resolution/Creation r=0.350) compared to mere co-presence in AI-involved threads. The central claim is that productive GenAI participation depends on these specific interaction-design elements rather than AI presence alone.

Significance. If the reported associations can be shown to reflect the design elements rather than selection, the work supplies concrete, replicable guidance for scaling GenAI facilitation in connectivist MOOCs. It credits the sizable sample and explicit reporting of correlation coefficients and p-values, which allow readers to assess effect magnitudes directly. The emphasis on reciprocal exchange and mandatory human review offers a falsifiable design hypothesis for future controlled studies.

major comments (1)
  1. [Iteration 2 results] Iteration 2 results (abstract and results section): the claim that direct learner-agent interaction produces higher presence (r=0.350 for Resolution/Creation, r=0.206 for Integration, r=0.186 for social presence) is load-bearing for the central design-dependence argument, yet the analysis relies on observed correlations without randomization, thread-level matching, or controls for learner motivation, prior engagement, or topic self-selection. This leaves the associations vulnerable to the alternative explanation that more motivated learners preferentially initiate or enter AI-visible threads.
minor comments (2)
  1. [Abstract] The abstract states N=606 but does not break out unique participants versus total enrollments or per-iteration sample sizes; adding these figures would clarify the effective power for the reported correlations.
  2. [Methods] The description of discourse-adaptive roles would benefit from a brief table or figure showing the exact decision rules or prompts used to classify responses as Guide versus Amplifier.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the thoughtful and detailed review. We address the single major comment on the Iteration 2 results below.

read point-by-point responses

  1. Referee: [Iteration 2 results] Iteration 2 results (abstract and results section): the claim that direct learner-agent interaction produces higher presence (r=0.350 for Resolution/Creation, r=0.206 for Integration, r=0.186 for social presence) is load-bearing for the central design-dependence argument, yet the analysis relies on observed correlations without randomization, thread-level matching, or controls for learner motivation, prior engagement, or topic self-selection. This leaves the associations vulnerable to the alternative explanation that more motivated learners preferentially initiate or enter AI-visible threads.

    Authors: We appreciate the referee's observation. The manuscript consistently describes the findings as associations (using the phrasing 'associated with') and does not claim that direct interaction 'produces' higher presence. Nevertheless, we agree that the observational design leaves the results open to selection bias, as more motivated or engaged learners may preferentially enter AI-visible threads or initiate direct interaction. In the revision we will (1) add an explicit paragraph in the Limitations section discussing this alternative explanation, (2) qualify the abstract and results language to stress the correlational nature of the evidence, and (3) outline how future randomized or matched-thread experiments could test the design elements more rigorously. These changes will be incorporated. revision: yes

Circularity Check

0 steps flagged

No significant circularity: empirical design-based study with observed correlations

full rationale

The paper reports results from a design-based research study involving iterative deployment of an AI facilitation workflow in a cMOOC, followed by measurement of correlations (e.g., r values for social and cognitive presence indicators) between interaction forms and presence metrics. No mathematical derivations, equations, fitted parameters renamed as predictions, or self-referential definitions appear in the provided text. The central claims rest on empirical observations across iterations rather than any reduction of outputs to inputs by construction. Self-citations, if present, are not load-bearing for the reported associations. This is self-contained empirical work with no evidence of the enumerated circularity patterns.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Relies on established Community of Inquiry framework for social and cognitive presence measures; no new free parameters, ad-hoc axioms, or invented entities are introduced.

axioms (1)
  • standard math Standard assumptions underlying Pearson correlation analysis (linearity, independence of observations)
    Invoked when reporting r and p values for presence measures.

pith-pipeline@v0.9.0 · 5672 in / 1110 out tokens · 60273 ms · 2026-05-14T00:03:19.989084+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

73 extracted references · 73 canonical work pages

  1. [1]

    Randy Garrison

    Zehra Akyol, Norm Vaughan, and D. Randy Garrison. 2011. The impact of course duration on the development of a community of inquiry.Interactive Learning Environments19, 3 (June 2011), 231–246. doi:10.1080/10494820902809147

    work page doi:10.1080/10494820902809147 2011

  2. [2]

    Terry Anderson and Jon Dron. 2011. Three generations of distance education pedagogy.The International Review of Research in Open and Distributed Learning 12, 3 (March 2011), 80. doi:10.19173/irrodl.v12i3.890

    work page doi:10.19173/irrodl.v12i3.890 2011

  3. [3]

    Terry Anderson and Julie Shattuck. 2012. Design-Based Research: A Decade of Progress in Education Research?Educational Researcher41, 1 (Jan. 2012), 16–25. doi:10.3102/0013189X11428813

    work page doi:10.3102/0013189x11428813 2012

  4. [4]

    Hemanth Asirvatham, Elliott Mokski, and Andrei Shleifer. 2026. GPT as a Measurement Tool. doi:10.3386/w34834

    work page doi:10.3386/w34834 2026

  5. [5]

    Shen Ba, Zachari Swiecki, Yuanru Tan, Guoqing Lu, David Williamson Shaffer, and Dragan Gašević. 2026. Community of inquiry in motion: Modeling inquiry dynamics with movement analysis (MOVA).Computers & Education242 (March 2026), 105513. doi:10.1016/j.compedu.2025.105513

    work page doi:10.1016/j.compedu.2025.105513 2026

  6. [6]

    font-variant:small-caps

    Shen Ba, Ying Zhan, Lingyun Huang, and Guoqing Lu. 2025. Investigating the impact of <span style="font-variant:small-caps;">ChatGPT</span> -assisted feedback on the dynamics and outcomes of online inquiry-based discussion. British Journal of Educational Technology56, 5 (Sept. 2025), 1710–1734. doi:10. 1111/bjet.13605

    work page 2025

  7. [7]

    Xuemei Bai and Xiaoqing Gu. 2025. How does peer-led teaching presence promote the development of cognitive presence? Evidence based on behavioural sequence analysis.Australasian Journal of Educational Technology41, 3 (May 2025), 26–44. doi:10.14742/ajet.10020

    work page doi:10.14742/ajet.10020 2025

  8. [8]

    Xuemei Bai, Xiaoqing Gu, and Rifa Guo. 2023. More factors, better understanding: model verification and construct validity study on the community of inquiry in MOOC.Education and Information Technologies28, 8 (Aug. 2023), 10483–10506. doi:10.1007/s10639-023-11604-z

    work page doi:10.1007/s10639-023-11604-z 2023

  9. [9]

    Yun-Qi Bai and Jian-Jun Xiao. 2023. The impact of cMOOC learners’ interaction on content production.Interactive Learning Environments31, 7 (Oct. 2023), 4464–4475. doi:10.1080/10494820.2021.1969955 15 citations (CrossRef 2026/1/31)

    work page doi:10.1080/10494820.2021.1969955 2023

  10. [10]

    Aicha Bakki, Lahcen Oubahssi, Sébastien George, and Chihab Cherkaoui. 2019. MOOCAT: A visual authoring tool in the cMOOC context.Education and Informa- tion Technologies24, 2 (March 2019), 1185–1209. doi:10.1007/s10639-018-9807-2

    work page doi:10.1007/s10639-018-9807-2 2019

  11. [11]

    Hamsa Bastani, Osbert Bastani, Alp Sungu, Haosen Ge, Özge Kabakcı, and Rei Mariman. 2025. Generative AI without guardrails can harm learning: Evidence from high school mathematics.Proceedings of the National Academy of Sciences 122, 26 (July 2025), e2422633122. doi:10.1073/pnas.2422633122

    work page doi:10.1073/pnas.2422633122 2025

  12. [12]

    Federico Battiston, Valerio Capraro, Fariba Karimi, Sune Lehmann, Andrea Bam- berg Migliano, Onkar Sadekar, Angel Sánchez, and Matjaž Perc. 2025. Higher- order interactions shape collective human behaviour.Nature Human Behaviour 9, 12 (Dec. 2025), 2441–2457. doi:10.1038/s41562-025-02373-5

    work page doi:10.1038/s41562-025-02373-5 2025

  13. [13]

    Daniela Castellanos-Reyes. 2021. The dynamics of a MOOC’s learner-learner interaction over time: A longitudinal network analysis.Computers in Human Behavior123 (Oct. 2021), 106880. doi:10.1016/j.chb.2021.106880

    work page doi:10.1016/j.chb.2021.106880 2021

  14. [14]

    Li Chen and Yaqian Xu. 2022. Theoretical Development of Connectivism through Innovative Application in China.Canadian Journal of Learning and Technology 48, 4 (Nov. 2022). doi:10.21432/cjlt28255 Number: 4

    work page doi:10.21432/cjlt28255 2022

  15. [15]

    internet + education

    Li Chen, Yaqian Xu, and Xinyi He. 2025. New concepts of knowledge, learn- ing, and curriculum for “internet + education”.Beijing International Review of Education7, 4 (Dec. 2025), 273–287. doi:10.1177/25902547251395881

    work page doi:10.1177/25902547251395881 2025

  16. [16]

    JinJu Duan and Qing Gao. 2025. Connectivist knowledge production and learning success in distributed social networks: structural equation modeling approach. Interactive Learning Environments(Sept. 2025), 1–19. doi:10.1080/10494820.2025. 2562172

    work page doi:10.1080/10494820.2025 2025

  17. [17]

    font- variant:small-caps

    Shihui Feng. 2025. Group interaction patterns in generative <span style="font- variant:small-caps;">AI</span> -supported collaborative problem solving: Net- work analysis of the interactions among students and a <span style="font- variant:small-caps;">GAI</span> chatbot.British Journal of Educational Tech- nology56, 5 (Sept. 2025), 2125–2145. doi:10.1111/...

    work page doi:10.1111/bjet.13611 2025

  18. [18]

    Ángel Fidalgo-Blanco, María Luisa Sein-Echaluce, and Francisco José García- Peñalvo. 2016. From massive access to cooperation: lessons learned and proven results of a hybrid xMOOC/cMOOC pedagogical approach to MOOCs.Interna- tional Journal of Educational Technology in Higher Education13, 1 (Dec. 2016), 24. doi:10.1186/s41239-016-0024-z

    work page doi:10.1186/s41239-016-0024-z 2016

  19. [19]

    Kevin Fuchs. 2023. Exploring the opportunities and challenges of NLP models in higher education: is Chat GPT a blessing or a curse?Frontiers in Education8 (2023). https://www.frontiersin.org/articles/10.3389/feduc.2023.1166682

    work page doi:10.3389/feduc.2023.1166682 2023

  20. [20]

    Ming Gao, Jingjing Zhang, and Jiang Zhang. 2025. Diversity of interactions within connectivist learning context: Insights from flow of collective attention. International Journal of Educational Technology in Higher Education22, 1 (Dec. 2025), 76. doi:10.1186/s41239-025-00575-5

    work page doi:10.1186/s41239-025-00575-5 2025

  21. [21]

    D.Randy Garrison, Terry Anderson, and Walter Archer. 1999. Critical Inquiry in a Text-Based Environment: Computer Conferencing in Higher Education.The Internet and Higher Education2, 2-3 (March 1999), 87–105. doi:10.1016/S1096- 7516(00)00016-6

    work page doi:10.1016/s1096- 1999

  22. [22]

    Randy Garrison, Terry Anderson, and Walter Archer

    D. Randy Garrison, Terry Anderson, and Walter Archer. 2001. Critical think- ing, cognitive presence, and computer conferencing in distance education. American Journal of Distance Education15, 1 (Jan. 2001), 7–23. doi:10.1080/ 08923640109527071

    work page 2001

  23. [23]

    Granovetter

    Mark S. Granovetter. 1973. The Strength of Weak Ties.Amer. J. Sociology78, 6 (1973), 1360–1380. https://www.jstor.org/stable/2776392

    work page arXiv 1973

  24. [24]

    Pengyue Guo, Nadira Saab, Lin Wu, and Wilfried Admiraal. 2021. The Community of Inquiry perspective on students’ social presence, cognitive presence, and academic performance in online project-based learning.Journal of Computer Assisted Learning37, 5 (Oct. 2021), 1479–1493. doi:10.1111/jcal.12586

    work page doi:10.1111/jcal.12586 2021

  25. [25]

    Songhee Han and Min Kyung Lee. 2022. FAQ chatbot and inclusive learning in massive open online courses.Computers & Education179 (2022), 104395. https://www.sciencedirect.com/science/article/pii/S0360131521002724

    work page 2022

  26. [26]

    Weijiao Huang and Khe Foon Hew. 2025. Facilitating Online Self-Regulated Learning and Social Presence Using Chatbots: Evidence-Based Design Principles. IEEE Transactions on Learning Technologies18 (2025), 56–71. doi:10.1109/TLT. 2024.3523199

    work page doi:10.1109/tlt 2025

  27. [27]

    Weijiao Huang, Khe Foon Hew, and Luke K. Fryer. 2022. Chatbots for language learning—Are they really useful? A systematic review of chatbot-supported language learning.Journal of Computer Assisted Learning38, 1 (Feb. 2022), 237–257. doi:10.1111/jcal.12610

    work page doi:10.1111/jcal.12610 2022

  28. [28]

    Lewis Johnson and James C

    W. Lewis Johnson and James C. Lester. 2018. Pedagogical Agents: Back to the Future.AI Magazine39, 2 (June 2018), 33–44. doi:10.1609/aimag.v39i2.2793

    work page doi:10.1609/aimag.v39i2.2793 2018

  29. [29]

    Graesser

    Srećko Joksimović, Nia Dowell, Oleksandra Poquet, Vitomir Kovanović, Dragan Gašević, Shane Dawson, and Arthur C. Graesser. 2018. Exploring development of social capital in a CMOOC through language and discourse.The Internet and Higher Education36 (Jan. 2018), 54–64. doi:10.1016/j.iheduc.2017.09.004

    work page doi:10.1016/j.iheduc.2017.09.004 2018

  30. [30]

    Enkelejda Kasneci, Kathrin Sessler, Stefan Küchemann, Maria Bannert, Daryna Dementieva, Frank Fischer, Urs Gasser, Georg Groh, Stephan Günnemann, Eyke Hüllermeier, Stephan Krusche, Gitta Kutyniok, Tilman Michaeli, Claudia Nerdel, Jürgen Pfeffer, Oleksandra Poquet, Michael Sailer, Albrecht Schmidt, Tina Sei- del, Matthias Stadler, Jochen Weller, Jochen Kuh...

    work page doi:10.1016/j.lindif.2023.102274 2023

  31. [31]

    c” and the “x

    Jeremy Knox. 2018. Beyond the “c” and the “x”: Learning with algorithms in massive open online courses (MOOCs).International Review of Education64, 2 (April 2018), 161–178. doi:10.1007/s11159-018-9707-0

    work page doi:10.1007/s11159-018-9707-0 2018

  32. [32]

    Xi Kong, Zhi Liu, Changsheng Chen, Sannyuya Liu, Zhenguo Xu, and Qianhui Tang. 2025. Exploratory study of an AI-supported discussion representational tool for online collaborative learning in a Chinese university.The Internet and Higher Education64 (Jan. 2025), 100973. doi:10.1016/j.iheduc.2024.100973

    work page doi:10.1016/j.iheduc.2024.100973 2025

  33. [33]

    Rita Kop, Hélène Fournier, and John Sui Fai Mak. 2011. A pedagogy of abundance or a pedagogy to support human beings? Participant support on massive open online courses.International Review of Research in Open and Distributed Learning 12, 7 (2011), 74–93

    work page 2011

  34. [35]

    Richardson

    Kadir Kozan and Jennifer C. Richardson. 2014. Interrelationships between and among social, teaching, and cognitive presence.The Internet and Higher Education 21 (April 2014), 68–73. doi:10.1016/j.iheduc.2013.10.007

    work page doi:10.1016/j.iheduc.2013.10.007 2014

  35. [37]

    Shuang Li, Xinyi He, and Jiaqi Chen. 2022. Exploring the relationship between interaction patterns and social capital accumulation in connectivist learning. Interactive Learning Environments0, 0 (Dec. 2022), 1–22. doi:10.1080/10494820. 2022.2157839

    work page doi:10.1080/10494820 2022

  36. [38]

    Zhi Liu, Huimin Duan, Shiqi Liu, Rui Mu, Sannyuya Liu, and Zongkai Yang. 2024. Improving knowledge gain and emotional experience in online learning with knowledge and emotional scaffolding-based conversational agent.Educational Technology & Society27, 2 (April 2024). doi:10.30191/ETS.202404_27(2).RP08

    work page doi:10.30191/ets.202404_27(2).rp08 2024

  37. [39]

    Zhi Liu, Xi Kong, Sannyuya Liu, Zongkai Yang, and Cuishuang Zhang. 2022. Looking at MOOC discussion data to uncover the relationship between discussion pacings, learners’ cognitive presence and learning achievements.Education and Information Technologies27, 6 (July 2022), 8265–8288. doi:10.1007/s10639-022- 10943-7

    work page doi:10.1007/s10639-022- 2022

  38. [40]

    Sriraam Natarajan, Saurabh Mathur, Sahil Sidheekh, Wolfgang Stammer, and Kristian Kersting. 2025. Human-in-the-loop or AI-in-the-loop? Automate or Collaborate?. InProceedings of the AAAI Conference on Artificial Intelligence, Vol. 39. 28594–28600. https://ojs.aaai.org/index.php/AAAI/article/view/35083

    work page 2025

  39. [41]

    Ha Nguyen. 2023. Role design considerations of conversational agents to facilitate discussion and systems thinking.Computers & Education192 (Jan. 2023), 104661. doi:10.1016/j.compedu.2022.104661 Xiao et al

    work page doi:10.1016/j.compedu.2022.104661 2023

  40. [42]

    Elvis Ortega-Ochoa, Marta Arguedas, and Thanasis Daradoumis. 2024. Empathic pedagogical conversational agents: A systematic literature review.British Journal of Educational Technology55, 3 (May 2024), 886–909. doi:10.1111/bjet.13413

    work page doi:10.1111/bjet.13413 2024

  41. [43]

    Sami Paavola, Lasse Lipponen, and Kai Hakkarainen. 2004. Models of Inno- vative Knowledge Communities and Three Metaphors of Learning.Review of Educational Research74, 4 (Dec. 2004), 557–576. doi:10.3102/00346543074004557

    work page doi:10.3102/00346543074004557 2004

  42. [44]

    Jason Phang, Michael Lampe, Lama Ahmad, Sandhini Agarwal, Cathy Mengying Fang, Auren R Liu, Valdemar Danry, Eunhae Lee, Samantha W T Chan, Pat Pataranutaporn, and Pattie Maes. 2025. Investigating affective use and emotional well-being on ChatGPT. (March 2025)

    work page 2025

  43. [45]

    Oleksandra Poquet, Vitomir Kovanović, Pieter de Vries, Thieme Hennis, Srećko Joksimović, Dragan Gašević, and Shane Dawson. 2018. Social presence in massive open online courses.International Review of Research in Open and Distributed Learning19, 3 (2018)

    work page 2018

  44. [46]

    Brenda Praggastis, Sinan Aksoy, Dustin Arendt, Mark Bonicillo, Cliff Joslyn, Emilie Purvine, Madelyn Shapiro, and Ji Young Yun. 2024. HyperNetX: A Python package for modeling complex network data as hypergraphs.Journal of Open Source Software9, 95 (March 2024), 6016. doi:10.21105/joss.06016

    work page doi:10.21105/joss.06016 2024

  45. [47]

    Sarah Prestridge, Katherine Main, and Mirjam Schmid. 2024. Identifying how classroom teachers develop presence online: breaking the fourth wall in online learning.Education and Information Technologies29, 2 (Feb. 2024), 1357–1377. doi:10.1007/s10639-023-11714-8

    work page doi:10.1007/s10639-023-11714-8 2024

  46. [48]

    THOMAS REEVES. 2006. Design research from a technology perspective. In Educational Design Research. Routledge. Num Pages: 15

    work page 2006

  47. [49]

    Richardson, Yukiko Maeda, Jing Lv, and Secil Caskurlu

    Jennifer C. Richardson, Yukiko Maeda, Jing Lv, and Secil Caskurlu. 2017. Social presence in relation to students’ satisfaction and learning in the online environ- ment: A meta-analysis.Computers in Human Behavior71 (June 2017), 402–417. doi:10.1016/j.chb.2017.02.001

    work page doi:10.1016/j.chb.2017.02.001 2017

  48. [50]

    Rivera, Mariane Frenay, Bernadette Charlier, and Valérie Swaen

    Dennis A. Rivera, Mariane Frenay, Bernadette Charlier, and Valérie Swaen. 2025. Houston, we’ve addressed a problem: a layer design for MOOC forums to improve navigation, participation, and interactions.Journal of Computing in Higher Education(April 2025). doi:10.1007/s12528-025-09436-6

    work page doi:10.1007/s12528-025-09436-6 2025

  49. [51]

    Vitor Rolim, Rafael Ferreira, Rafael Dueire Lins, and Dragan Gˇasević. 2019. A network-based analytic approach to uncovering the relationship between social and cognitive presences in communities of inquiry.The Internet and Higher Education42 (July 2019), 53–65. doi:10.1016/j.iheduc.2019.05.001

    work page doi:10.1016/j.iheduc.2019.05.001 2019

  50. [52]

    Zaffar Ahmed Shaikh and Shakeel Ahmed Khoja. 2012. Role of Teacher in Personal Learning Environments.Digital Education Review(June 2012). https: //eric.ed.gov/?id=EJ972714 ERIC Number: EJ972714

    work page 2012

  51. [53]

    Peter Shea, Suzanne Hayes, Jason Vickers, Mary Gozza-Cohen, Sedef Uzuner, Ruchi Mehta, Anna Valchova, and Prahalad Rangan. 2010. A re-examination of the community of inquiry framework: Social network and content analysis. The Internet and Higher Education13, 1-2 (Jan. 2010), 10–21. doi:10.1016/j.iheduc. 2009.11.002

    work page doi:10.1016/j.iheduc 2010

  52. [54]

    Peter Shea, Jennifer Richardson, and Karen Swan. 2022. Building bridges to advance the Community of Inquiry framework for online learning.Educational Psychologist57, 3 (July 2022), 148–161. doi:10.1080/00461520.2022.2089989

    work page doi:10.1080/00461520.2022.2089989 2022

  53. [55]

    G. Siemens. 2005. Connectivism: A learning theory for the digital age.In- ternational Journal of Instructional Technology and Distance Learning2, 1 (2005). https://www.scopus.com/inward/record.uri?eid=2-s2.0-33745587105& partnerID=40&md5=63cd1564999432bc8cc6c44bcbde840e tex.publication_stage: Final

    work page 2005

  54. [56]

    Yu Song, Longchao Huang, Lanqin Zheng, Mengya Fan, and Zehao Liu. 2025. Interactions with generative AI chatbots: unveiling dialogic dynamics, students’ perceptions, and practical competencies in creative problem-solving.Interna- tional Journal of Educational Technology in Higher Education22, 1 (March 2025),

    work page 2025

  55. [57]

    doi:10.1186/s41239-025-00508-2

    work page doi:10.1186/s41239-025-00508-2

  56. [58]

    Songhee Han and Min Liu. 2024. Equity at the forefront: A systematic research and development process of chatbot curriculum for massive open online courses. Educational Technology & Society27, 4 (Oct. 2024), 339–351. doi:10.30191/ETS. 202410_27(4).SP08

    work page doi:10.30191/ets 2024

  57. [59]

    Louise Stoll, Ray Bolam, Agnes McMahon, Mike Wallace, and Sally Thomas

    work page

  58. [60]

    2006), 221–258

    Professional Learning Communities: A Review of the Literature.Journal of Educational Change7, 4 (Dec. 2006), 221–258. doi:10.1007/s10833-006-0001-8

    work page doi:10.1007/s10833-006-0001-8 2006

  59. [61]

    Yulin Tian and Jianjun Xiao. 2026. The measurement and characteristic analysis of learner interaction levels in cMOOCs based on path analysis.Interactive Learning Environments34, 1 (Jan. 2026), 297–319. doi:10.1080/10494820.2025.2498534

    work page doi:10.1080/10494820.2025.2498534 2026

  60. [62]

    Voogt, H

    J. Voogt, H. Westbroek, A. Handelzalts, A. Walraven, S. McKenney, J. Pieters, and B. de Vries. 2011. Teacher learning in collaborative curriculum design.Teaching and Teacher Education27, 8 (Nov. 2011), 1235–1244. doi:10.1016/j.tate.2011.07.003

    work page doi:10.1016/j.tate.2011.07.003 2011

  61. [63]

    Cixiao Wang and Jianjun Xiao. 2024. Who will participate in online collabo- rative problem solving? A longitudinal network analysis.Interactive Learning Environments32, 10 (Nov. 2024), 6534–6551. doi:10.1080/10494820.2023.2266523

    work page doi:10.1080/10494820.2023.2266523 2024

  62. [64]

    Wal- lace, and Linlin Li

    Xinghua Wang, Qian Liu, Hui Pang, Seng Chee Tan, Jun Lei, Matthew P. Wal- lace, and Linlin Li. 2023. What matters in AI-supported learning: A study of human-AI interactions in language learning using cluster analysis and epis- temic network analysis.Computers & Education194 (March 2023), 104703. doi:10.1016/j.compedu.2022.104703

    work page doi:10.1016/j.compedu.2022.104703 2023

  63. [65]

    Zhijun Wang, Terry Anderson, Li Chen, and Elena Barbera. 2017. Interaction pattern analysis in cMOOCs based on the connectivist interaction and engage- ment framework.British Journal of Educational Technology48, 2 (March 2017), 683–699. doi:10.1111/bjet.12433

    work page doi:10.1111/bjet.12433 2017

  64. [66]

    Joshua Weidlich, Derya Orhan Göksün, and Karel Kreijns. 2023. Extending social presence theory: social presence divergence and interaction integration in online distance learning.Journal of Computing in Higher Education35, 3 (Dec. 2023), 391–412. doi:10.1007/s12528-022-09325-2

    work page doi:10.1007/s12528-022-09325-2 2023

  65. [67]

    Alyssa Friend Wise and Yi Cui. 2018. Learning communities in the crowd: Characteristics of content related interactions and social relationships in MOOC discussion forums.Computers & Education122 (July 2018), 221–242. doi:10.1016/ j.compedu.2018.03.021

    work page 2018

  66. [68]

    Jianjun Xiao. 2026. Exploring interaction patterns in open learning environments: integrating network dynamics and cognitive engagement.Interactive Learning Environments0, 0 (Feb. 2026), 1–19. doi:10.1080/10494820.2026.2632766 _eprint: https://www.tandfonline.com/doi/pdf/10.1080/10494820.2026.2632766

    work page doi:10.1080/10494820.2026.2632766 2026

  67. [69]

    Jianjun Xiao, Cixiao Wang, and Wenmei Zhang. 2025. Modeling Collaborative Problem Solving Dynamics from Group Discourse: A Text-Mining Approach with Synergy Degree Model. doi:10.1145/3785022.3785049 arXiv:2512.13061 [cs]

    work page doi:10.1145/3785022.3785049 2025

  68. [70]

    Wanli Xing, Taehyun Kim, Yukyeong Song, Hai Li, Chenglu Li, and Jinhee Kim

    work page

  69. [71]

    Unveiling interaction patterns between students and generative AI teach- able agent: Focusing on students’ agency and AI agents’ authority.British Journal of Educational Technologyn/a, n/a (Feb. 2026). doi:10.1111/bjet.70038 _eprint: https://bera-journals.onlinelibrary.wiley.com/doi/pdf/10.1111/bjet.70038

    work page doi:10.1111/bjet.70038 2026

  70. [72]

    Yaqian Xu and Yang Yang. 2024. Research on the patterns of the organic growth knowledge production based on a dual perspective of content and behaviour. British Journal of Educational Technology55, 3 (May 2024), 1145–1166. doi:10. 1111/bjet.13418

    work page 2024

  71. [73]

    Lixiang Yan, Samuel Greiff, Ziwen Teuber, and Dragan Gašević. 2024. Promises and challenges of generative artificial intelligence for human learning.Nature Human Behaviour8, 10 (Oct. 2024), 1839–1850. doi:10.1038/s41562-024-02004-5

    work page doi:10.1038/s41562-024-02004-5 2024

  72. [74]

    Wenmei Zhang and Cixiao Wang. 2023. Comparative interaction patterns of groups in an open network environment: The role of facilitators in collaborative learning.Journal of Computer Assisted Learning40, 1 (Sept. 2023), 136–157. doi:10.1111/jcal.12873

    work page doi:10.1111/jcal.12873 2023

  73. [75]

    Wenting Zou, Xiao Hu, Zilong Pan, Chenglu Li, Ying Cai, and Min Liu. 2021. Exploring the relationship between social presence and learners’ prestige in MOOC discussion forums using automated content analysis and social network analysis.Computers in Human Behavior115 (Feb. 2021), 106582. doi:10.1016/j. chb.2020.106582

    work page doi:10.1016/j 2021