Defining AI companions: a research agenda-from artificial companions for learning to general artificial companions for Global Harwell
The DOI matches its Crossref record; 14 of the 14 references it cites were found in a registry or an index.
Abstract
This paper examines and outlines the future exploration of AI companions, focusing on three main areas. Firstly, it defines AI companions as sophisticated AI entities designed for supporting and enhancing human experiences in daily activities, such as learning, working, and others. They encompass emotional, social, and practical aspects of daily life while fostering interactions and relationships with humans. Secondly, the paper provides a historical review of AI companions for learning (i.e., AI learning companions), offering an overview of their conceptualization, development, and utilization in educational contexts and insights into their potential future trajectory. Lastly, a research agenda is presented, which includes AI learning companions—outlining key questions, challenges, and goals for integrating AI companions into learning environments—and the pursuit of Global ‘Harwell’ (a portmanteau of ‘harmony’ and ‘wellbeing’). This suggests that, in addition to transforming education, AI companions can contribute to individual wellbeing as well as broader humanitarian objectives. Beyond contributing to economic growth and efficiency, it is fundamentally important to address the most pressing global challenges or human crises of our time, foster understanding and cooperation among researchers and practitioners in different fields, and hence pave the way toward a future world marked by Global Harwell. To further advocate and demonstrate the pivotal role of AI companions in achieving these genuine objectives, we propose the General Artificial Companion Hypothesis. All these endeavors, however, begin with AI companions for learning. As Mandela said, “Education is the most powerful weapon which you can use to change the world.”
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Chou et al. (2025) argue that this represents a broader and significant historical shift in human-technology relationships: from tools, to assistants, to companions.”
From Artificial intelligence and unconditional love: the rise of generative AI as an alternative form of mental health support (2026) · cites this paper as Chou et al. (2025)
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Self-Regulated Learning and Academic Achievement: An Overview
Zimmerman, B. J. (1990) · Educational Psychologist · doi:10.1207/s15326985ep2501_2 · Cited by 3,687
Unifying Large Language Models and Knowledge Graphs: A Roadmap
Pan, S., Luo, L., Wang, Y., Chen, C., Wang, J., & Wu, X. (2024) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2024.3352100 · Cited by 1,376 · cited 1×
“…retrieval-augmented generation (RAG) combines the strengths of LLMs and information retrieval systems to provide highly accurate and contextually relevant answers (Pan et al., 2024), enabling ALCs to offer students precise and comprehensive information throughout their learning process.”
BackgroundMoreno, R., Mayer, R. E., Spires, H. A., & Lester, J. C. (2001) · Cognition and Instruction · doi:10.1207/s1532690xci1902_02 · Cited by 912
Chatbots applications in education: A systematic review
Okonkwo, C. W., & Ade-Ibijola, A. (2021) · Computers and Education: Artificial Intelligence · doi:10.1016/j.caeai.2021.100033 · Cited by 879
Huang, W., Hew, K. F., & Fryer, L. K. (2022) · Journal of Computer Assisted Learning · doi:10.1111/jcal.12610 · Cited by 830 · cited 1×
“Chatbot companions utilize natural language processing techniques to communicate with students in natural language (e.g., Huang et al., 2022; Huang et al., 2008; Hwang & Chang, 2023; Kim et al., 2022; Kuhail et al., 2023; Liu et al., 2024; Okonkwo & Ade-Ibijola, 2021; Ruan et al., 2019; Skjuve et…”
BackgroundAutoTutor: an intelligent tutoring system with mixed-initiative dialogue.
Graesser, A. C., Chipman, P., Haynes, B. C., & Olney, A. (2005) · IEEE Transactions on Education · doi:10.1109/te.2005.856149 · Cited by 750
A review of opportunities and challenges of chatbots in education
Hwang, G.-J., & Chang, C.-Y. (2023) · Interactive Learning Environments · doi:10.1080/10494820.2021.1952615 · Cited by 530 · cited 1×
“…utilize natural language processing techniques to communicate with students in natural language (e.g., Huang et al., 2022; Huang et al., 2008; Hwang & Chang, 2023; Kim et al., 2022; Kuhail et al., 2023; Liu et al., 2024; Okonkwo & Ade-Ibijola, 2021; Ruan et al., 2019; Skjuve et al., 2021).”
BackgroundA metacognitive view of individual differences in self-regulated learning
Winne, P. H. (1996) · Learning and Individual Differences · doi:10.1016/s1041-6080(96)90022-9 · Cited by 486 · cited 1×
“…learning behaviors, such as overestimated self-assessments, inappropriate goal-setting, lack of regulation, and avoidant help-seeking, which can lead to poor learning performance (e.g., Chen et al., 2019; Chou et al., 2015; 2018; Chou & Zou, 2020; Winne, 1996; Zimmerman & Schunk, 1989).”
BackgroundCreating Large Language Model Applications Utilizing LangChain: A Primer on Developing LLM Apps Fast
Topsakal, O., & Akinci, T. C. (2023) · International Conference on Applied Engineering and Natural Sciences · doi:10.59287/icaens.1127 · Cited by 331 · cited 1×
“Moreover, large language models (LLMs) like GPT-4 enhance the ability of generative AI to understand and produce human-like text, which can be utilized in tutoring systems, essay feedback, and language translation tools (Bozkurt, 2023; Topsakal & Akinci, 2023).”
BackgroundThe promise and challenges of multimodal learning analytics
Cukurova, M., Giannakos, M., & Martinez‐Maldonado, R. (2020) · British Journal of Educational Technology · doi:10.1111/bjet.13015 · Cited by 169 · cited 1×
“Multimodal learning analytics provide a clearer picture of student’s learning processes and statuses, and they bring more accurate predictions through the translation, coordination, alignment or fusion of multimodal data than a single data source (Baltrušaitis et al., 2019; Cukurova et al., 2020).”
Method
Cited by
Liu, T., Lo, T.-Y., Wen, K.-H., Sun, Y., & Wei, Z.-Q. (2026) · Frontiers in Psychology · doi:10.3389/fpsyg.2025.1687686 · Cited by 13
Miller, V., Hill, M. J., & Moreira, T. (2026) · Communication and Change · doi:10.1007/s44382-026-00036-1 · Cited by 2
Metadata
- Authors
- Chou, C.-Y., Chan, T.-W., Chen, Z.-H., Liao, C.-Y., Shih, J.-L., Wu, Y.-T., Chang, B., Yeh, C. Y. C., Hung, H.-C., & Cheng, H.
- Venue
- Research and Practice in Technology Enhanced Learning
- Published
- 2024-10-22
- Volume
- 20
- Pages
- 032
- ISSN
- 1793-7078
- Cited by
- 27
- References
- 14
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Defining AI companions: a research agenda-from artificial companions for learning to general artificial companions for Global Harwell. https://app.cochat.ai/papers/10.58459/rptel.2025.20032
- The paper
- Chou, C.-Y., Chan, T.-W., Chen, Z.-H., Liao, C.-Y., Shih, J.-L., Wu, Y.-T., Chang, B., Yeh, C. Y. C., Hung, H.-C., & Cheng, H. (2024). Defining AI companions: a research agenda-from artificial companions for learning to general artificial companions for Global Harwell. Research and Practice in Technology Enhanced Learning. https://doi.org/10.58459/rptel.2025.20032
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.