Pith. sign in

REVIEW 4 major objections 6 minor 79 references

Conversational AI as a Catalyst for Informal Learning: An Empirical Large-Scale Study on LLM Use in Everyday Learning

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper claims that large language models have already become deeply embedded tools in everyday informal learning, with 88% of a 776-person German adult sample using them for learning and four distinct learner profiles emerging from…

desk verdict A solid large-scale descriptive survey whose four-learner typology needs LCA robustness reporting before it can be taken as settled. read the letter →

arxiv 2506.11789 v1 pith:FRVNPRBA submitted 2025-06-13 cs.HC

Na{j̣a Terzimehić , Babette Bühler , Enkelejda Kasneci This is my paper
classification cs.HC
keywords informallearninglargelanguagemodelseverydaylatentclassanalysislearnerprofilesChatGPTsurveystudyhuman-computerinteraction
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper claims that large language models have already become ordinary tools in everyday informal learning, not emerging technologies: 88% of a 776-person German adult sample report using them for learning, and the study identifies four distinct learner profiles based on how, when, and where they use LLMs. The authors argue that adoption is driven less by AI knowledge than by curiosity and affinity for technology, while mistrust of accuracy and privacy concerns are the main barriers. They also document a paradox: learners use LLMs to fact-check information even while doubting their accuracy, and report low privacy concern while taking few protective measures. A sympathetic reader would take away that LLM-based informal learning has moved into the mainstream and that design should address different learner types rather than a one-size-fits-all assistant.

What carries the argument

Latent class analysis (LCA), a probabilistic method that assumes an unobserved categorical variable explains associations among observed indicators, applied to three indicator sets: devices used (desktop, laptop, smartphone, tablet), learning contexts (higher education, K-12, lifelong learning, professional development), and learning tasks (summarizing, brainstorming, problem-solving, and related activities). The four-class solution was chosen via AIC and BIC, and the class-conditional probabilities define the profiles. This machinery carries the argument by turning self-reported usage patterns into distinct learner types that organize the paper's design implications.

What would settle it

Re-run the latent class analysis on an independent sample using the same indicators: if information criteria favor a different number of classes, or if class membership probabilities are low, the four-profile typology collapses. More directly, a representative household sample that finds everyday LLM learning adoption well below 88% would refute the claim that LLMs are already deeply embedded in everyday learning.

Watch

Extended reading notes

Core claim

The central discovery is that LLMs are already deeply embedded in everyday learning routines: 88% of the surveyed adults use them for learning, and latent class analysis of their device use, learning contexts, and tasks yields four stable learner profiles — Structured Knowledge Builders, Self-Guided Explorers, Analytical Problem Solvers, and Adaptive Power Users. Demographic adoption gaps mirror early technology adoption (younger, male, more educated), yet once people start learning with LLMs, the depth and style of engagement is shaped more by perceived effectiveness and technology affinity than by gender. The paper also finds that learners simultaneously mistrust LLM accuracy and rely on it for fact-checking, and that they do not perceive themselves as over-reliant, complicating the overreliance narrative.

Load-bearing premise

The four-type typology rests on the assumption that latent class analysis uncovers real, stable subgroups from the chosen indicators; the paper selects a four-class solution without reporting the fit of alternative solutions or replication checks, so the existence of exactly four profiles is not independently verified.

Editorial extensions

If this is right

  • If 88% adoption is accurate, LLMs are now a mainstream informal learning infrastructure, and design efforts should focus on improving embedded tools rather than driving adoption.
  • The four profiles imply that one-size-fits-all interfaces fail distinct user groups, so adaptive or tailored designs could better support each type.
  • The fact-checking/accuracy paradox suggests that providing sources and confidence indicators may be more effective than simply warning about hallucinations.
  • Since learners do not perceive themselves as over-reliant, interventions to mitigate overreliance should target actual behavior rather than self-reported attitudes.
  • The gender gap in adoption but not in engagement depth suggests that education and exposure, not capability, may be the lever for closing the gap.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the four profiles replicate in other populations, they could become a standard segmentation for LLM learning design; a testable extension is to see whether the profiles predict actual learning outcomes rather than self-reports.
  • The 88% figure likely overstates the general population because the sample was drawn from an online research platform with tech-savvy participants; a representative household sample would probably show lower but still substantial adoption.
  • The paradox of fact-checking despite distrust suggests a 'convenience over accuracy' tradeoff: on-demand access wins even when users know about hallucinations, which predicts that adding source citations may increase trust but not substantially change usage.
  • Future work could test whether nudging Self-Guided Explorers toward more structured reflection improves actual learning gains compared with letting their mobile, on-the-go pattern persist.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. This paper reports a survey of 776 German Prolific participants about their use of large language models (LLMs) for everyday informal learning. The authors find that 88% of respondents use LLMs for learning, describe demographic and motivational correlates of adoption, and use latent class analysis (LCA) on devices, contexts, and tasks to identify four learner profiles (Structured Knowledge Builders, Self-Guided Explorers, Analytical Problem Solvers, Adaptive Power Users). They then relate these profiles to perceived effectiveness, overreliance, and privacy attitudes, and offer design implications for LLM-based learning tools. The paper positions itself as a large-scale, mixed-method complement to earlier qualitative and student-specific studies.

Significance. If the descriptive findings hold, the paper provides a valuable snapshot of informal LLM learning in an early-adopter population, with a large sample and a broad set of measures. The proposed four-type typology could be generative for HCI research on personalized learning interfaces, and the paper explicitly acknowledges several limitations (e.g., Prolific sample, self-report, ChatGPT-assisted questionnaire design) while offering concrete design implications. The main analytic contribution, the LCA-based typology, currently rests on insufficiently reported model-selection evidence, and one of the headline 'paradoxical behavior' claims is internally inconsistent. With the LCA diagnostics and consistency issues addressed, the empirical contribution would be solid and useful to the community.

major comments (4)
  1. [§3.3.1 and §4.3.3 and Appendix B] The selection of the four-class LCA solution is not independently verifiable. The text reports only the chosen solution's log-likelihood (–10575.98), AIC (21397.96), and BIC (21953.82) and refers to 'Figure B' for comparisons, but Appendix B contains no such table or figure. Without fit statistics for the 2-, 3-, 5-, and 6-class models, entropy values, number of random starts, or any replication (bootstrap or split-half), the claim that the four-class solution is optimal is unsupported. The assertion that 'modal posterior probabilities confirmed the robustness of the solution' (§4.3.3) is given without numeric support. Because the four class shares are nearly equal (23.8%–25.7%) and the profiles in §5.4.3 and the regression in §4.4.5 all depend on this solution, the authors should provide the full model-comparison table, entropy and average posterior probabilities per class, and a robustness check (e.g., different starts or split-half).
  2. [§5.2 and §4.3.1] The fact-checking paradox is based on an unverifiable number. Section 5.2 states that '301 learners in our sample suggest using LLMs to verify information,' but §4.3.1 reports the relevant task frequencies as Explanatory Inquiries (n=471), Factual Queries (n=429), and Tutorial Requests (n=414), and no 'fact-checking' item or n=301 appears in §4.3.1 or in Figure 5. Either the item and its count need to be reported in §4.3.1, or the claim about fact-checking being the most common use must be revised. As written, the central 'paradoxical behavior' in the abstract and §5.2 rests on an internal inconsistency.
  3. [§4.1 and Tables 1–2] The RQ1 and RQ2 comparisons involving the 'Non-LLM Users' group (n=15) are given more weight than the sample supports. For example, the statement that 87% of non-users earn below EUR50K is based on 13 people, and the 'mistrust' barrier is based on n=6; these percentages have very wide confidence intervals. The paper should either restrict claims about non-users to descriptive counts with explicit uncertainty, or drop the n=15 from inferential comparisons. This is particularly relevant because the abstract frames the paper around 'who remains hesitant.'
  4. [§4.4.5] The multinomial logistic regression treats the LCA class assignments as known, ignoring classification error. With modal assignment, misclassification tends to bias coefficients toward zero and can distort the pattern of predictors. Since the LCA classes are themselves the dependent variable, the analysis should either use a one-step (simultaneous) latent class regression or incorporate posterior probabilities (e.g., via multiple imputation of class membership). At minimum, the paper should report average posterior probabilities to allow readers to gauge classification uncertainty.
minor comments (6)
  1. [Abstract and §4] The abstract states '88% of our respondents' while §4 reports 678 respondents (87%) out of 776 (87.4%); use consistent rounding throughout.
  2. [§4.2] The text says 'one eight' instead of 'one-eighth', and the challenges list reports both 12% (n=82) and 9% (n=82) for different items; one of these counts must be wrong since 12% and 9% of 678 differ.
  3. [§3.3.1] The reference to 'as depicted in Figure B' is misleading because Appendix B is empty after its title; either include the promised fit table or remove the reference.
  4. [§5.4.3] The phrase 'Self-guided Explorers and Analytical Power Users' appears to be a typo for 'Self-Guided Explorers and Adaptive Power Users'.
  5. [§4.3.3] The profile name 'Academic Knowledge Builders' is used once in the final paragraph of this section, while the rest of the paper uses 'Structured Knowledge Builders'; use one consistent label.
  6. [§4.1] The ANOVA and chi-square tests would benefit from effect sizes (e.g., η², Cramér's V) and a clear statement of which groups are included in each comparison (especially whether the n=15 non-users are excluded).

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the survey findings and latent class profiles are grounded in new self-report data, and the few self-citations are background constructs, not load-bearing evidence.

full rationale

This paper is an empirical survey, so the central claims (88% adoption, demographic differences, motivating and discouraging factors, four learner profiles) are summaries of newly collected self-report data rather than derivations from assumptions. The latent class analysis estimates classes from participants' reported devices, contexts, and tasks; the four-class solution is selected by AIC/BIC, and while the manuscript omits important diagnostics (no fit values for alternative class counts, no entropy, no replication checks), this is a robustness and reporting limitation, not circularity, because the classes are not defined in terms of the paper's conclusions. The cited prior works with overlapping authorship (e.g., the inquiry taxonomy in [7] and the educational opportunities/challenges discussion in [38]) are used as measurement instruments and background framing, not as evidence that forces the empirical results. No uniqueness theorem is imported from the authors' prior work, and no fitted parameter is renamed as a prediction. The limitations section acknowledges generalizability constraints (e.g., German Prolific sample, English fluency, possible social desirability bias), which are validity concerns rather than circular reasoning. Accordingly, no specific circular step can be quoted or exhibited, and the appropriate finding is no significant circularity with a minimal score reflecting only the presence of minor, non-load-bearing self-citations.

Assumptions & free parameters 3 free parameters · 3 assumptions · 1 invented entities

The main modeling choices are the number of latent classes and the indicator set for the LCA; the main domain assumptions are that self-reports reflect behavior, that the Prolific sample generalizes, and that the LCA's categorical latent variable assumption holds. The four learner profiles are data-derived constructs without external validation.

free parameters (3)
  • Number of latent classes in LCA = 4
    Selected based on AIC and BIC as the 'best balance', but no other class solutions, entropy, or bootstrap indices are reported, so the choice cannot be independently evaluated.
  • LCA indicator set = devices, learning contexts, learning tasks
    These three categorical dimensions were chosen by the authors; different or additional indicators could produce a different typology.
  • Attention check exclusion rule = exclusion of participants who marked a task as used but later gave frequency as 'never used'
    This post-hoc rule removed 25 participants; its validity is assumed and not validated against an external criterion.
assumptions (3)
  • domain assumption LCA assumes that an underlying latent categorical variable accounts for the associations among observed variables.
    Invoked in Section 3.3.1; if the true structure is continuous, the four-class solution may be an artifact of the model.
  • domain assumption Self-reported LLM use and task frequencies accurately reflect actual learning behavior.
    The survey relies entirely on self-report; attention checks filter some inconsistent responses but cannot validate ground truth (Section 3.2).
  • domain assumption The German Prolific sample with fluent English is adequate to draw conclusions about everyday LLM learning adoption.
    Stated in Section 3.2 and acknowledged as a limitation in Section 5.5; if the sample is unrepresentative, the adoption rate and demographic effects do not generalize.
invented entities (1)
  • Four learner profiles (Structured Knowledge Builders, Self-Guided Explorers, Analytical Problem Solvers, Adaptive Power Users)
    purpose: Describe latent classes of LLM learners based on device, context, and task indicators.
    Derived from LCA on this sample; no external validation or replication is provided, so they are not established beyond this study.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Conversational AI as a Catalyst for Informal Learning: An Empirical Large-Scale Study on LLM Use in Everyday Learning." pith.science (2026). https://pith.science/paper/FRVNPRBA

@misc{pith2026250611789,
  author       = {Pith},
  title        = {Pith review of: Conversational AI as a Catalyst for Informal Learning: An Empirical Large-Scale Study on LLM Use in Everyday Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FRVNPRBA}},
  note         = {Machine review of arXiv:2506.11789}
}
read the original abstract

Large language models have not only captivated the public imagination but have also sparked a profound rethinking of how we learn. In the third year following the breakthrough launch of ChatGPT, everyday informal learning has been transformed as diverse user groups explore these novel tools. Who is embracing LLMs for self-directed learning, and who remains hesitant? What are their reasons for adoption or avoidance? What learning patterns emerge with this novel technological landscape? We present an in-depth analysis from a large-scale survey of 776 participants, showcasing that 88% of our respondents already incorporate LLMs into their everyday learning routines for a wide variety of (learning) tasks. Young adults are at the forefront of adopting LLMs, primarily to enhance their learning experiences independently of time and space. Four types of learners emerge across learning contexts, depending on the tasks they perform with LLMs and the devices they use to access them. Interestingly, our respondents exhibit paradoxical behaviours regarding their trust in LLMs' accuracy and privacy protection measures. Our implications emphasize the importance of including different media types for learning, enabling collaborative learning, providing sources and meeting the needs of different types of learners and learning by design.

Figures

Figures reproduced from arXiv: 2506.11789 by the authors.

Figure 1
Figure 1. Demographic distributions: (a) age and (b) gender [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 2
Figure 2. Demographic distributions: (a) income and (b) education level [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Distributions of ATI scores and personality traits: (a) ATI score distribution, (b) big-5 personality distribution [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figures from the paper (10 more)
Figure 4
Figure 4. Figure 4: Proportion of reported (a) challenges and (b) benefits experienced by learners when using LLMs for learning [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]
Figure 5
Figure 5. Figure 5: Frequency of mentioned inquiries Following this, we zoomed into the learning scenario and inquired about the learning tasks participants complete with the assistance of LLMs, presenting them with a multiple-choice list of options. Participants could then report the fre…
Figure 6
Figure 6. Figure 6: Frequency of the learning tasks participants perform with LLMs [PITH_FULL_IMAGE:figures/full_fig_p015_6.png]
Figure 7
Figure 7. Figure 7: Participants’ hourly distribution of LLM use for learning over a 24-hour period [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]
Figure 8
Figure 8. Figure 8: Frequency of the mentioned a) learning contexts with LLMs b) devices to access LLMs for learning and c) platforms to access [PITH_FULL_IMAGE:figures/full_fig_p017_8.png]
Figure 9
Figure 9. Figure 9: Latent Class Analysis: Survey Item Probabilities by Learner Profile: Grouped by Profile [PITH_FULL_IMAGE:figures/full_fig_p018_9.png]
Figure 10
Figure 10. Figure 10: Latent Class Analysis: Survey Item Probabilities by Learner Profile: Grouped by Survey Item [PITH_FULL_IMAGE:figures/full_fig_p019_10.png]
Figure 11
Figure 11. Figure 11: Likert-scale distribution of LLM perceptions of over-reliance and usefulness [PITH_FULL_IMAGE:figures/full_fig_p020_11.png]
Figure 12
Figure 12. Figure 12: Likert-scale distribution of LLM perceptions on productivity (above) and privacy concerns (below) [PITH_FULL_IMAGE:figures/full_fig_p021_12.png]
Figure 13
Figure 13. Figure 13: LLM learners’ responses on a) how likely they are to continue using LLMs for learning and b) whether they would recommend [PITH_FULL_IMAGE:figures/full_fig_p022_13.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

79 extracted references · 50 canonical work pages

  1. [1]

    Ghadah Al Murshidi, Galina Shulgina, Anastasiia Kapuza, and Jamie Costley. 2024. How understanding the limitations and risks of using ChatGPT can contribute to willingness to use. Smart Learning Environments 11, 1 (2024), 36

    work page 2024

  2. [2]

    Yazid Albadarin, Mohammed Saqr, Nicolas Pope, and Markku Tukiainen. 2024. A systematic literature review of empirical research on ChatGPT in education. Discover Education 3, 1 (26 May 2024), 60. https://doi.org/10.1007/s44217-024-00138-2

    work page doi:10.1007/s44217-024-00138-2 2024

  3. [3]

    Riggs, Maryam Esmaeili, Rastin Mastali Majdabadkohne, and Morteza Pasehvar

    Aram Bahrini, Mohammadsadra Khamoshifar, Hossein Abbasimehr, Robert J. Riggs, Maryam Esmaeili, Rastin Mastali Majdabadkohne, and Morteza Pasehvar. 2023. ChatGPT: Applications, Opportunities, and Threats. In 2023 Systems and Information Engineering Design Symposium (SIEDS) . 274–279. https://doi.org/10.1109/SIEDS58326.2023.10137850

    arXiv 2023

  4. [4]

    Baker and Aaron Hawn

    Ryan S. Baker and Aaron Hawn. 2022. Algorithmic Bias in Education. International Journal of Artificial Intelligence in Education 32, 4 (01 Dec 2022), 1052–1092. https://doi.org/10.1007/s40593-021-00285-9

    doi:10.1007/s40593-021-00285-9 2022

  5. [5]

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

    Emily M. Bender, Timnit Gebru, Angelina McMillan-Major, and Shmargaret Shmitchell. 2021. On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?. In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (Virtual Event, Canada) (FAccT ’21). Association for Computing Machinery, New York, NY, USA, 610–623. https:/...

    arXiv 2021

  6. [6]

    Leonardo Berti, Flavio Giorgi, and Gjergji Kasneci. 2025. Emergent Abilities in Large Language Models: A Survey. arXiv:2503.05788 [cs.LG] https://arxiv.org/abs/2503.05788

    arXiv 2025

  7. [7]

    Anna Bodonhelyi, Efe Bozkir, Shuo Yang, Enkelejda Kasneci, and Gjergji Kasneci. 2024. User Intent Recognition and Satisfaction with Large Language Models: A User Study with ChatGPT. arXiv:2402.02136 [cs.HC] https://arxiv.org/abs/2402.02136

    arXiv 2024

  8. [8]

    Zhihui Cai, Xitao Fan, and Jianxia Du. 2017. Gender and attitudes toward technology use: A meta-analysis. Computers and Education 105 (2017), 1–13. https://doi.org/10.1016/j.compedu.2016.11.003

    work page doi:10.1016/j.compedu.2016.11.003 2017

Show all 79 references
  1. [9]

    Callanan, C

    M. Callanan, C. Cervantes, and M. Loomis. 2011. Informal learning. Wiley interdisciplinary reviews. Cognitive science 2 6 (2011), 646–655. https://doi.org/10.1002/wcs.143

    2011 doi

  2. [10]

    Cerasoli, G

    Christopher P. Cerasoli, G. Alliger, Jamie S. Donsbach, J. Mathieu, S. Tannenbaum, and Karin A. Orvis. 2018. Antecedents and Outcomes of Informal Learning Behaviors: a Meta-Analysis. Journal of Business and Psychology 33 (2018), 203–230. https://doi.org/10.1007/S10869-017-9492-Y

    2018 doi

  3. [11]

    Cecilia Ka Yuk Chan and Wenxin Zhou. 2023. An expectancy value theory (EVT) based instrument for measuring student perceptions of generative AI. Smart Learning Environments 10, 1 (2023), 64

    2023

  4. [12]

    Chiu, Qi Xia, Xinyan Zhou, Ching Sing Chai, and Miaoting Cheng

    Thomas K.F. Chiu, Qi Xia, Xinyan Zhou, Ching Sing Chai, and Miaoting Cheng. 2023. Systematic literature review on opportunities, challenges, and future research recommendations of artificial intelligence in education. Computers and Education: Artificial Intelligence 4 (2023), ...

    2023

  5. [13]

    Fabio Clarizia, Francesco Colace, Marco Lombardi, Francesco Pascale, and Domenico Santaniello. 2018. Chatbot: An Education Support System for Student. In Cyberspace Safety and Security , Arcangelo Castiglione, Florin Pop, Massimo Ficco, and Francesco Palmieri (Eds.). Springer ...

    2018

  6. [14]

    Ali Darvishi, Hassan Khosravi, Shazia Sadiq, Dragan Gašević, and George Siemens. 2024. Impact of AI assistance on student agency. Computers and Education 210 (2024), 104967. https://doi.org/10.1016/j.compedu.2023.104967

    2024

  7. [15]

    PHILLIPS DAVISON

    W. PHILLIPS DAVISON. 1983. The Third-Person Effect in Communication. Public Opinion Quarterly 47, 1 (01 1983), 1–15. https://doi.org/10.1086/ 268763 arXiv:https://academic.oup.com/poq/article-pdf/47/1/1/5382397/47-1-1.pdf

    1983

  8. [16]

    Edward L Deci and Richard M Ryan. 2012. Self-determination theory. Handbook of theories of social psychology 1, 20 (2012), 416–436

    2012

  9. [17]

    Isaac Dinner, Eric J Johnson, Daniel G Goldstein, and Kaiya Liu. 2011. Partitioning default effects: why people choose not to choose. Journal of Experimental Psychology: Applied 17, 4 (2011), 332

    2011

  10. [18]

    Fiona Draxler. 2023. Designing intelligent support for learning from and in everyday contexts. http://nbn-resolving.de/urn:nbn:de:bvb:19-325783

    2023

  11. [19]

    Fiona Draxler, Daniel Buschek, Mikke Tavast, Perttu Hämäläinen, Albrecht Schmidt, Juhi Kulshrestha, and Robin Welsch. 2023. Gender, Age, and Technology Education Influence the Adoption and Appropriation of LLMs. arXiv:2310.06556 [cs.CY] https://arxiv.org/abs/2310.06556

    2023 arXiv

  12. [20]

    So what if ChatGPT wrote it?

    Yogesh K. Dwivedi, Nir Kshetri, Laurie Hughes, Emma Louise Slade, Anand Jeyaraj, Arpan Kumar Kar, Abdullah M. Baabdullah, Alex Koohang, Vishnupriya Raghavan, Manju Ahuja, Hanaa Albanna, Mousa Ahmad Albashrawi, Adil S. Al-Busaidi, Janarthanan Balakrishnan, Yves Barlette, Sripar...

    2023

  13. [21]

    Michael Eraut. 2004. Informal learning in the workplace. Studies in Continuing Education 26, 2 (2004), 247–273. https://doi.org/10.1080/ 158037042000225245 arXiv:https://doi.org/10.1080/158037042000225245

    2004 doi

  14. [22]

    Haim Eshach. 2007. Bridging In-school and Out-of-school Learning: Formal, Non-Formal, and Informal Education. Journal of Science Education and Technology 16, 2 (01 Apr 2007), 171–190. https://doi.org/10.1007/s10956-006-9027-1

    2007 doi

  15. [23]

    Justin R Evans, Matt Karlsven, Spencer B Perry, and R Kimmons. 2020. Informal learning. The Students’ Guide to Learning Design and Research; Kimmons, R., Ed (2020)

    2020

  16. [24]

    Martin Flintham, Christian Karner, Khaled Bachour, Helen Creswick, Neha Gupta, and Stuart Moran. 2018. Falling for Fake News: Investigating the Consumption of News via Social Media. In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems (Montreal QC, C...

    2018

  17. [25]

    Göran Folkestad. 2006. Formal and informal learning situations or practices vs formal and informal ways of learning. British Journal of Music Education 23, 2 (2006), 135–145. https://doi.org/10.1017/S0265051706006887

    2006 doi

  18. [26]

    Tim Fütterer, Christian Fischer, Anastasiia Alekseeva, Xiaobin Chen, Tamara Tate, Mark Warschauer, and Peter Gerjets. 2023. ChatGPT in education: global reactions to AI innovations. Scientific Reports 13, 1 (15 Sep 2023), 15310. https://doi.org/10.1038/s41598-023-42227-6

    2023 doi

  19. [27]

    Xiao Ge, Chunchen Xu, Daigo Misaki, Hazel Rose Markus, and Jeanne L Tsai. 2024. How Culture Shapes What People Want From AI. In Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems (Honolulu, HI, USA) (CHI ’24). Association for Computing Machinery, New ...

    2024

  20. [28]

    Simone Grassini. 2023. Shaping the Future of Education: Exploring the Potential and Consequences of AI and ChatGPT in Educational Settings. Education Sciences 13, 7 (2023). https://doi.org/10.3390/educsci13070692

    2023 doi

  21. [29]

    Kwon, Ahmed Hosny Saleh Metwally, Ahmed Tlili, Simone Bassanelli, Antonio Bucchiarone, Sujit Gujar, Lennart E

    Reza Hadi Mogavi, Chao Deng, Justin Juho Kim, Pengyuan Zhou, Young D. Kwon, Ahmed Hosny Saleh Metwally, Ahmed Tlili, Simone Bassanelli, Antonio Bucchiarone, Sujit Gujar, Lennart E. Nacke, and Pan Hui. 2024. ChatGPT in education: A blessing or a curse? A qualitative study explo...

    2024

  22. [30]

    Juho Hamari, Nicolai Hanner, and Jonna Koivisto. 2017. Service quality explains why people use freemium services but not if they go premium: An empirical study in free-to-play games. Int. J. Inf. Manag. 37 (2017), 1449–1459. https://doi.org/10.1016/j.ijinfomgt.2016.09.004

    2017 doi

  23. [31]

    Why pay premium in freemium services?

    Juho Hamari, Nicolai Hanner, and Jonna Koivisto. 2020. "Why pay premium in freemium services?" A study on perceived value, continued use and purchase intentions in free-to-play games. International Journal of Information Management 51 (2020), 102040. https://doi.org/10.1016/j....

    2020

  24. [32]

    Muhammad Hamdan. 2023. The Impact of ChatGPT on Students: Positive Effects and Potential Negative Effects — moderndiplomacy.eu. https: //moderndiplomacy.eu/2023/03/14/the-impact-of-chatgpt-on-students-positive-effects-and-potential-negative-effects/. [Accessed 04-04-2025]

    2023

  25. [33]

    Marie Hornberger, Arne Bewersdorff, and Claudia Nerdel. 2023. What do university students know about Artificial Intelligence? Development and validation of an AI literacy test. Computers and Education: Artificial Intelligence 5 (2023), 100165. https://doi.org/10.1016/j.caeai.2...

    2023

  26. [34]

    Mohammad Hosseini, Catherine A Gao, David M Liebovitz, Alexandre M Carvalho, Faraz S Ahmad, Yuan Luo, Ngan MacDonald, Kristi L Holmes, and Abel Kho. 2023. An exploratory survey about using ChatGPT in education, healthcare, and research. Plos one 18, 10 (2023), e0292216

    2023

  27. [35]

    Lei Huang, Weijiang Yu, Weitao Ma, Weihong Zhong, Zhangyin Feng, Haotian Wang, Qianglong Chen, Weihua Peng, Xiaocheng Feng, Bing Qin, and Ting Liu. 2025. A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions. ACM Trans. Inf. S...

    2025 doi

  28. [36]

    Maurice Jakesch, Zana Buçinca, Saleema Amershi, and Alexandra Olteanu. 2022. How Different Groups Prioritize Ethical Values for Responsible AI. In Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency (Seoul, Republic of Korea) (FAccT ’22). Assoc...

    2022

  29. [37]

    Celina Kacperski, Denis Bonnay, Juhi Kulshrestha, Peter Selb, Andreas Spitz, and Roberto Ulloa. 2024. Characteristics of ChatGPT users from Germany: implications for the digital divide from web tracking data. arXiv:2309.02142 [cs.CY] https://arxiv.org/abs/2309.02142

    2024 arXiv

  30. [38]

    Enkelejda Kasneci, Kathrin Seßler, Stefan Küchemann, Maria Bannert, Daryna Dementieva, Frank Fischer, Urs Gasser, Georg Groh, Stephan Günnemann, Eyke Hüllermeier, et al. 2023. ChatGPT for good? On opportunities and challenges of large language models for education. Learning an...

    2023

  31. [39]

    Kasthuri and S

    E. Kasthuri and S. Balaji. 2021. A Chatbot for Changing Lifestyle in Education. In 2021 Third International Conference on Intelligent Communication Technologies and Virtual Mobile Networks (ICICV) . 1317–1322. https://doi.org/10.1109/ICICV50876.2021.9388633

    2021

  32. [40]

    Nils Knoth, Antonia Tolzin, Andreas Janson, and Jan Marco Leimeister. 2024. AI literacy and its implications for prompt engineering strategies. Computers and Education: Artificial Intelligence 6 (2024), 100225. https://doi.org/10.1016/j.caeai.2024.100225

    2024

  33. [41]

    Robert B Kozma. 1991. Learning with media. Review of educational research 61, 2 (1991), 179–211

    1991

  34. [42]

    Justin Kruger and David Dunning. 1999. Unskilled and unaware of it: how difficulties in recognizing one’s own incompetence lead to inflated self-assessments. Journal of personality and social psychology 77, 6 (1999), 1121. 2025-08-09 01:06. Page 28 of 1–32. An Empirical Large-...

    1999

  35. [43]

    Ka Hei Carrie Lau, Bhada Yun, Samuel Saruba, Efe Bozkir, and Enkelejda Kasneci. 2024. Wrapped in Anansi’s Web: Unweaving the Impacts of Generative-AI Personalization and VR Immersion in Oral Storytelling. arXiv preprint arXiv:2409.16894 (2024)

    2024 arXiv

  36. [44]

    Daniel Lee and Edward Palmer. 2025. Prompt engineering in higher education: a systematic review to help inform curricula. International Journal of Educational Technology in Higher Education 22, 1 (2025), 7

    2025

  37. [45]

    Marina Lepp and Joosep Kaimre. 2025. Does generative AI help in learning programming: Students’ perceptions, reported use and relation to performance. Computers in Human Behavior Reports 18 (2025), 100642. https://doi.org/10.1016/j.chbr.2025.100642

    2025

  38. [46]

    Zhehui Liao, Maria Antoniak, Inyoung Cheong, Evie Yu-Yen Cheng, Ai-Heng Lee, Kyle Lo, Joseph Chee Chang, and Amy X. Zhang. 2024. LLMs as Research Tools: A Large Scale Survey of Researchers’ Usage and Perceptions. arXiv:2411.05025 [cs.CL] https://arxiv.org/abs/2411.05025

    2024 arXiv

  39. [47]

    Xi Lin. 2024. Exploring the Role of ChatGPT as a Facilitator for Motivating Self-Directed Learning Among Adult Learners. Adult Learning 35, 3 (2024), 156–166. https://doi.org/10.1177/10451595231184928 arXiv:https://doi.org/10.1177/10451595231184928

    2024 doi

  40. [48]

    Xinru Lin and Luyang Li. 2025. Implicit Bias in LLMs: A Survey. arXiv:2503.02776 [cs.CL] https://arxiv.org/abs/2503.02776

    2025 arXiv

  41. [49]

    D. W. LIVINGSTONE. 2006. INFORMAL LEARNING: CONCEPTUAL DISTINCTIONS AND PRELIMINARY FINDINGS. Counterpoints 249 (2006), 203–227. http://www.jstor.org/stable/42979596

    2006

  42. [50]

    Wenhan Lyu, Yimeng Wang, Tingting (Rachel) Chung, Yifan Sun, and Yixuan Zhang. 2024. Evaluating the Effectiveness of LLMs in Introductory Computer Science Education: A Semester-Long Field Study. In Proceedings of the Eleventh ACM Conference on Learning @ Scale (Atlanta, GA, US...

    2024

  43. [51]

    Janice Malcolm, Phil Hodkinson, and Helen Colley. 2003. The interrelationships between informal and formal learning. Journal of workplace learning 15, 7/8 (2003), 313–318

    2003

  44. [52]

    Amelia Manuti, Serafina Pastore, Anna Fausta Scardigno, Maria Luisa Giancaspro, and Daniele Morciano. 2015. Formal and informal learning in the workplace: A research review. International journal of training and development 19, 1 (2015), 1–17

    2015

  45. [53]

    Marsick and Marie Volpe

    Victoria J. Marsick and Marie Volpe. 1999. The Nature and Need for Informal Learning. Advances in Developing Human Resources 1, 3 (1999), 1–9. https://doi.org/10.1177/152342239900100302 arXiv:https://doi.org/10.1177/152342239900100302

    1999 doi

  46. [54]

    Richard E. Mayer. 2009. Multimedia Learning (2 ed.). Cambridge University Press

    2009

  47. [55]

    Meredith Ringel Morris. 2023. Scientists’ Perspectives on the Potential for Generative AI in their Fields. arXiv:2304.01420 [cs.CY] https: //arxiv.org/abs/2304.01420

    2023 arXiv

  48. [56]

    Ahnaf Chowdhury Niloy, Md Ashraful Bari, Jakia Sultana, Rup Chowdhury, Fareha Meem Raisa, Afsana Islam, Saadman Mahmud, Iffat Jahan, Moumita Sarkar, Salma Akter, Nurunnahar Nishat, Muslima Afroz, Amit Sen, Tasnem Islam, Mehedi Hasan Tareq, and Md Amjad Hossen. 2024. Why do stu...

    2024

  49. [57]

    Timothy J Nokes-Malach, J Elizabeth Richey, and Soniya Gadgil. 2015. When is it better to learn together? Insights from research on collaborative learning. Educational Psychology Review 27 (2015), 645–656

    2015

  50. [58]

    Rock Yuren Pang, Hope Schroeder, Kynnedy Simone Smith, Solon Barocas, Ziang Xiao, Emily Tseng, and Danielle Bragg. 2025. Understanding the LLM-ification of CHI: Unpacking the Impact of LLMs at CHI through a Systematic Literature Review. arXiv:2501.12557 [cs.HC] https: //arxiv....

    2025 arXiv

  51. [59]

    Minju Park, Sojung Kim, Seunghyun Lee, Soonwoo Kwon, and Kyuseok Kim. 2024. Empowering Personalized Learning through a Conversation-based Tutoring System with Student Modeling. In Extended Abstracts of the CHI Conference on Human Factors in Computing Systems (Honolulu, HI, USA...

    2024

  52. [60]

    Nermin Punar Özçelik and Gonca Yangın Ekşi. 2024. Cultivating writing skills: the role of ChatGPT as a learning assistant—a case study. Smart Learning Environments 11, 1 (18 Mar 2024), 10. https://doi.org/10.1186/s40561-024-00296-8

    2024 doi

  53. [61]

    Yao Qu, Michelle Xin Yi Tan, and Jue Wang. 2024. Disciplinary differences in undergraduate students’ engagement with generative artificial intelligence. Smart Learning Environments 11, 1 (2024), 51

    2024

  54. [62]

    Beatrice Rammstedt and Oliver P. John. 2007. Measuring personality in one minute or less: A 10-item short version of the Big Five Inventory in English and German. Journal of Research in Personality 41, 1 (2007), 203–212. https://doi.org/10.1016/j.jrp.2006.02.001

    2007 doi

  55. [63]

    Rogoff, M

    B. Rogoff, M. Callanan, Kris D. Gutiérrez, and F. Erickson. 2016. The Organization of Informal Learning. Review of Research in Education 40 (2016), 356 – 401. https://doi.org/10.3102/0091732X16680994

    2016 doi

  56. [64]

    Daniel Schugurensky. 2000. The forms of informal learning: Towards a conceptualization of the field

    2000

  57. [65]

    Kathrin Seßler, Tao Xiang, Lukas Bogenrieder, and Enkelejda Kasneci. 2023. PEER: empowering writing with large language models. In European conference on technology enhanced learning . Springer, 755–761

    2023

  58. [66]

    Farhana Shahid, Srujana Kamath, Annie Sidotam, Vivian Jiang, Alexa Batino, and Aditya Vashistha. 2022. ”It Matches My Worldview”: Examining Perceptions and Attitudes Around Fake Videos. In Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems (New Orlean...

    2022

  59. [67]

    Marita Skjuve, Asbjørn Følstad, and Petter Bae Brandtzaeg. 2023. The User Experience of ChatGPT: Findings from a Questionnaire Study of Early Users. In Proceedings of the 5th International Conference on Conversational User Interfaces (Eindhoven, Netherlands) (CUI ’23). Associa...

    2023

  60. [68]

    Matthias Stadler, Maria Bannert, and Michael Sailer. 2024. Cognitive ease at a cost: LLMs reduce mental effort but compromise depth in student scientific inquiry. Computers in Human Behavior 160 (2024), 108386. https://doi.org/10.1016/j.chb.2024.108386 2025-08-09 01:06. Page 2...

    2024

  61. [69]

    Jan-Philipp Stein, Tanja Messingschlager, Timo Gnambs, Fabian Hutmacher, and Markus Appel. 2024. Attitudes towards AI: measurement and associations with personality. Scientific Reports 14, 1 (05 Feb 2024), 2909. https://doi.org/10.1038/s41598-024-53335-2

    2024 doi

  62. [71]

    Artur Strzelecki. 2024. To use or not to use ChatGPT in higher education? A study of students’ acceptance and use of technology. Interactive learning environments 32, 9 (2024), 5142–5155

    2024

  63. [72]

    Jérémie Sublime and Ilaria Renna. 2024. Is ChatGPT Massively Used by Students Nowadays? A Survey on the Use of Large Language Models such as ChatGPT in Educational Settings. arXiv preprint arXiv:2412.17486 (2024)

    2024 arXiv

  64. [73]

    European Social Survey. 2025. ESS Data Portal | European Social Survey 2025. https://doi.org/10.21338/ess5e03_5

    2025 doi

  65. [74]

    Christiane Attig Thomas Franke and Daniel Wessel. 2019. A Personal Resource for Technology Interaction: Development and Validation of the Affinity for Technology Interaction (ATI) Scale. International Journal of Human–Computer Interaction 35, 6 (2019), 456–467. https://doi.org...

    2019

  66. [75]

    Chandan Kumar Tiwari, Mohd Abass Bhat, Shagufta Tariq Khan, Rajaswaminathan Subramaniam, and Mohammad Atif Irshad Khan. 2024. What drives students toward ChatGPT? An investigation of the factors influencing adoption and usage of ChatGPT. Interactive Technology and Smart Educat...

    2024

  67. [76]

    Hickey, Ronghuai Huang, and Brighter Agyemang

    Ahmed Tlili, Boulus Shehata, Michael Agyemang Adarkwah, Aras Bozkurt, Daniel T. Hickey, Ronghuai Huang, and Brighter Agyemang. 2023. What if the devil is my guardian angel: ChatGPT as a case study of using chatbots in education. Smart Learning Environments 10, 1 (22 Feb 2023),...

    2023 doi

  68. [77]

    Chaoran Wang, Zixi Li, and Curtis Bonk. 2024. Understanding self-directed learning in AI-Assisted writing: A mixed methods study of postsecondary learners. Computers and Education: Artificial Intelligence 6 (2024), 100247. https://doi.org/10.1016/j.caeai.2024.100247

    2024

  69. [78]

    Mo Wang, Minjuan Wang, Xin Xu, Lanqing Yang, Dunbo Cai, and Minghao Yin. 2024. Unleashing ChatGPT’s Power: A Case Study on Optimizing Information Retrieval in Flipped Classrooms via Prompt Engineering. IEEE Transactions on Learning Technologies 17 (2024), 629–641. https: //doi...

    2024

  70. [79]

    Yuxia Wang, Minghan Wang, Muhammad Arslan Manzoor, Fei Liu, Georgi Georgiev, Rocktim Jyoti Das, and Preslav Nakov. 2024. Factuality of Large Language Models: A Survey. arXiv:2402.02420 [cs.CL] https://arxiv.org/abs/2402.02420

    2024 arXiv

  71. [80]

    Ruben Weijers, Gabrielle Fidelis de Castilho, Jean-François Godbout, Reihaneh Rabbany, and Kellin Pelrine. 2024. Quantifying learning-style adaptation in effectiveness of LLM teaching. In Proceedings of the 1st Workshop on Personalization of Generative AI Systems (PERSONALIZE ...

    2024

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.