Introduction
The integration of active learning strategies with AI technologies in the development of OER represents an emerging approach to advancing open education in higher education, by promoting more accessible and responsive learning environments. On the one hand, AI technologies are redefining the educational experience by empowering dynamic, interactive, and personalized learning opportunities that extend beyond traditional classroom boundaries. On the other hand, the strategic guidance provided by educators complements these technological advances by ensuring that innovative practices are anchored to pedagogical principles.
Throughout this paper, the researcher examines a case study within a virtual social innovation laboratory in which participants collaborated to create a chatbot as an OER. The innovation element was the integration of AI into the OER by providing adaptive and responsive content, highlighting its defining features and development processes. It was intentionally designed to support the Human Resource Management (HRM) department in a private university. The educational foundation informing this active learning practice was mentoring, to guide the group dynamics, decision-making, and the integration of ethical considerations into the final product.
Therefore, the objective is to examine how mentoring, framed as an active learning strategy, contributes to the collaborative development of a generative AI-based OER within a virtual social innovation laboratory. It is guided by three Research Questions (RQ): 1) How do mentoring interventions support the development of a generative AI-based OER within a virtual laboratory environment? 2) How do mentoring practices influence group dynamics, decision-making, and the collaborative development of a generative AI-based OER? and 3) What are the characteristics of the generative AI-based OER developed by participants?
Active learning in digital environments
Active learning has been widely recognized as a pedagogical approach that promotes deep engagement, critical thinking, and the construction of knowledge. Yet, it is worth noting that active learning strategies are often described instead of defined in literature. Many definitions focus on social interaction (group work, discussions) or critical thinking (problem based – learning, simulations) (Doolittle, Wojdak & Walters 2023), while others emphasize a student-centered approach for continuous participation and reflection through motivating and challenging activities designed to deepen knowledge, develop searching, analyzing, and synthesizing skills, and encourage active adaptation to problem-solving (IFE 2024).
Moreover, when active learning is focused on the process, then it is defined as a form of learning in which students develop artifacts and receive formative assessment in several iterative steps (active – constructive – interactive), work in groups (cooperative), and connect learning with materials outside the classroom (reflective, authentic) (Jahnke et al. 2020), in such a way that they have more control and effective participation in the process, requiring multiple mental actions and constructions (Neves, Lima & Mesquita 2021).
In digital environments, active learning expands through the use of collaborative technologies, virtual simulations, and interactive platforms that offer opportunities for dynamic engagement and real-time feedback (Astudillo, Nogueira & Ortiz 2020; Cakula 2021), as video conference, cloud collaboration, web hosting and visual design have increased student engagement, promoted active involvement, and enhanced learning outcomes (Balalle 2024). Accordingly, these interactive and immersive tools, considered advanced OER, can be used to effectively support the development of required competencies in higher education (Meng 2023).
As students must learn to navigate diverse viewpoints and reach consensus on solutions (Martínez & Gómez 2025), the role of the educator in scaffolding and guiding students through these processes is often overlooked. The following are some perspectives that address the contribution of mentors.
The professor’s main role is organizing the educational process around active learning methods to guide students in formulating solutions, analyzing sources of information, researching issues, communicating effectively, analyzing the results and fostering engagement and deeper understanding (Víscu 2024). In an action learning session, “the coach helps learners stay focused on the problem through asking questions that reframe the problem or facilitate reflection” (Ferguson et al. 2019: 27). In virtual educational environments, professors act as mediators and motivators, integrating digital resources to enhance engagement and facilitate knowledge construction (Flores-Rivera & Meléndez-Tamayo 2024).
Nonetheless, it is important to recognize the “complexity of mentoring roles and the range of capabilities required to carry out the role effectively” (Thornton 2025: 366). Professors may require guidance or support in implementing strategies to enhance student engagement and collaboration, on available resources or training for optimal utilization of the platforms for active learning (Lando & Bowdler 2024).
Mentoring as a strategy to achieve shared goals
Mentoring plays a crucial role in guiding educational processes by offering structured support and expertise to learners (Horváth 2025). As an interpersonal pedagogical approach, mentoring bridges gaps between theoretical concepts and practical applications, contributing significantly to skill development and knowledge acquisition. It facilitates active learning by promoting reflection, autonomy, and supporting learners through complex tasks (Raptis et al. 2025).
Moreover, mentoring is a process where the mentors “demonstrate a range of cognitive coaching competencies … posing carefully constructed questions to stimulate reflection, paraphrasing, and using data to improve teaching and learning (Kamarudin et al. 2020: 291). Coaching, as a form of mentoring, implies a dynamic interpersonal relationship wherein an experienced individual offers technical expertise, advice, support, and inspiration to guide the learning process (IFE 2024). Studies on mentorship in online environments highlight the importance of structured interventions to replicate real-world dynamics, allowing mentors and mentees to connect and collaborate effectively in a learning partnership (Larsen et al. 2024).
Mentoring and coaching, while distinct, are complementary practices that support both personal and professional growth, and although this paper does not seek to establish a definitive distinction between mentoring and coaching, it acknowledges the need for further conceptual work to clarify their roles at the intersection of academic learning and professional training. For the purpose of this study, the term mentoring will be used in alignment with the social and educational innovation laboratories.
A mentorship framework
Following the exploration of general mentoring principles, the review now turns to the model that anchors the exploration of this study: the Cultural and Social Youth Entrepreneurship (CASYE) framework (Figure 1), an adaptable guide for entrepreneurial and innovation processes. Unlike traditional academic mentoring approaches, the CASYE framework is rooted in a co-creative and practice-oriented philosophy that emphasizes flexible, relational guidance over linear instruction (Diesis Network 2022a). As described in its handbook, “mentoring is important for supporting, encouraging, and guiding people to understand their full potential, and learning more from experienced professionals” (IARS et al. 2022: 5).
Figure 1
CASYE mentoring framework.
Source: Diesis Network (2022a), CASYE Mentoring Model Programme Framework.
The framework is composed of modular tools, reflective activities, and mentoring guides that structure the relationship around four developmental areas: idea development, hard skills, soft skills, and kick off (launch). This configuration makes CASYE particularly compatible with the participatory, progressive refinement approach of social and educational innovation laboratories.
Although references to the CASYE framework have been developed through an Erasmus+ initiative and are yet to be documented in academic publications, it is supported by a coherent body of structured outputs, including a background study examining the social economy landscape across Spain, Italy, Belgium, and the UK; a collection of 15 case studies highlighting good practices; and a mentoring model program framework tailored to youth organizations working with individuals facing social, economic, or geographical barriers (Diesis Network 2022b). These resources offer the conceptual and practical foundations of the CASYE framework, positioning it as a valuable, practice-oriented contribution to emerging literature on inclusive mentoring and social entrepreneurship.
Open Educational Resources and Artificial Intelligence
OER are “learning, teaching and research materials in any format and medium that reside in the public domain or are under copyright that have been released under an open license, that permit no-cost access, re-use, re-purpose, adaptation and redistribution by others” (UNESCO 2022: 4). According to Ramírez-Montoya (2020: 172) “with the OER, it is possible to create opportunities for constructing educational practices with great reach.” Numerous stakeholders have contributed to the advancement of OER by documenting their use, value, challenges and potential in educational contexts. Platforms such as EDUCAUSE (2025), OER Commons (2025), Open Education Global (2025), and The OER Knowledge Cloud (2025) provide open-access repositories, policy reports, and empirical studies in the subject. Although this is just a selection of key contributors, they exemplify the wider landscape of actors committed to promoting openness in education.
Moreover, the Dubai Declaration on OER (UNESCO 2024) highlights that emerging technologies, particularly artificial intelligence, have the potential to transform educational practices by reducing barriers to access and supporting global knowledge-sharing. This development presents both new opportunities and complex challenges for expanding the reach and impact of OER through AI integration.
AI, particularly generative, refers to AI systems capable of producing novel outputs, such as text, images, and audio, based on patterns learned from large datasets (OpenAI 2024). In education “these technologies hold the capacity to make learning more responsive, personalized, and scalable (learners’ developmental zones)” (CUAED 2023:62).
The integration of generative AI into OER is still an emerging field with a growing body of scholarly work. The development of AI-based OERs demonstrates the possibility for generative AI technologies to bring innovation into educational practices as well as for continuous professional development (Bosch & Kruger 2024). According to the Organisation for Economic Co-operation and Development, it also introduces critical challenges that must be addressed to align AI development with ethical and human-centered goals and ensure its responsible and equitable use (OECD 2024). Some examples of relevant initiatives include a policy observatory on AI (OECD 2025), research collaboration groups GAIA-GEN (UNAM 2023), and the Framework for Accessible and Equitable Artificial Intelligence in Education (Inclusive Design Research Centre 2024).
Social and educational innovation laboratories
Social and educational innovation laboratories are collaborative and experimental spaces designed to foster the collective and participatory construction of knowledge to solve the real-life problems of companies, government (including society) or both (Cortés, Ramírez, and Molina 2020). These laboratories “offer opportunities for higher education to work closely with professional practice with the emphasis on innovation research in real life” (Schuurman, De Marez & Ballon 2022: 30). They place a strong emphasis on informed learning and information literacy by integrating insights from science, interprofessional education, and community development, creating opportunities for individuals to connect, share expertise, and work toward a common goal through interdisciplinarity, active participation, and experimentation (Hughes, Foth & Mallan 2019; Yañez-Figueroa, Ramírez-Montoya & García-Peñalvo 2020).
When digitally mediated, social and educational innovation laboratories are spaces in which learners co-design solutions to real-world problems. They emphasize openness, participation, and collaboration to tackle societal challenges (Baran 2020; Gómez Zermeño & Alemán de la Garza 2021; Schrammel & Marschalek 2024; Valenzuela-Zubiaur et al. 2021). Furthermore, social innovation laboratories serve as digitally mediated spaces for active learning as participants “have enough knowledge of a problem area… to help each other… offer various approaches… and peer knowledge is promoted over expert knowledge… solving complex issues together and…develop problem-solving skills” (Fernández-Morales 2023: 38).
Methodology
This study adopted a qualitative case study design to examine mentoring dynamics and the collaborative development of an AI-based OER within a virtual social innovation laboratory. Case study is ideal for exploring “an event, a program or an activity or more than one individual” in depth (Creswell & Poth 2018: 104). The online context provided the conditions for tracing how mentoring strategies intersect within active learning, guided by digital ethnography, “adapting the traditional, in-person ethnographic research techniques to the study of online cultures and communities formed through computer-mediated communications” (Borkovich 2022: 1).
This methodology considers the researcher’s direct participation as a mentor, which provides an insider perspective on participants’ processes, interactions, and collaborative dynamics. The reflexive introspective ethnography of personal digital interactions (Hassan 2024) was appropriate to examine the dual role.
Context and participants
The laboratory was conducted Monday to Friday, fully online for a period of two weeks. It included two synchronous mentoring sessions per week and three group work sessions, supported by Zoom for meetings, Google Docs for content drafting and coordination, and WhatsApp for real-time communication. Materials containing information related to OER development were available on Moodle. The lab was framed within a project-based learning methodology and emphasized interdisciplinary collaboration, active learning, and social innovation. Mentoring was integrated as a pedagogical strategy to guide team processes and product development.
While several groups participated in the virtual laboratory and mentors were assigned by the lab organizers, this study concentrates on a single group directly mentored by the researcher. It included 11 participants, four women and seven men, all affiliated with the same university. Participants were professors and academic professionals with backgrounds in education, psychology, health sciences, communication, and digital technologies. Some had previously developed OER. According to the lab’s structure, two members acted as project promoters while the other nine contributed as collaborators. The mentor was affiliated to a different institution and was introduced to the group during the first session.
Data collection, processing and analysis
The sources for data collection were the researcher’s notes including reflections on mentoring interventions, observations of participants’ engagement, notes on incidents and challenges; and the ‘Lab notebook’ entries containing project documentation registered by the participants on the development processes of the AI-based OER. This approach to data collection is grounded in online fieldwork, social life and digitally mediated communication that engage with participants and gather their perspectives (Jensen et al. 2022).
Data processing started with an anonymization protocol to protect participants’ privacy. All names, email addresses, institutional affiliations and other personal identifiers were removed and replaced with alphanumeric data according to their roles (Prom1-promoter 1, Col2-collaborator2).
An inductive approach was applied to the mentor’s notes captured during Zoom meetings to identify emergent codes without imposing predefined categories. Then, those codes were deductively compared and reorganized around the CASYE framework, as a structured lens to examine mentoring interventions systematically. A second analysis was applied to map the collaborative actions recorded in the Lab notebook to the CASYE framework in search of connections with the mentor interventions. Finally, the AI-OER was analyzed according to UNESCO’s OER principles for universal access. This triangulation supported credibility and reduced bias.
As access to the digital field work was relatively short, relevant quotes were manually extracted from raw data and organized in Excel. After that, OpenAI ChatGPT model o3 was used for refining concepts, drafting, comparing and summarizing clean data. AI outputs were verified, edited, or discarded as necessary through an iterative review process. Decisions and interpretations were based on the researcher’s judgment.
Results and discussion
This section presents the results followed by the discussion on how structured mentoring facilitated the co-creation of an AI-OER around the research questions.
How do mentoring interventions support the development of a generative AI-based OER within a virtual laboratory environment?
Mentoring helped shape the project from its earliest stages by guiding decisions, autonomy, and thoughtful reflection, while the entrepreneurial-oriented structure of the CASYE framework was appropriate to examine mentoring within the nature of the social and innovation laboratories. Table 1 summarizes how interventions during the Mentoring approach and Idea development phases helped establish a shared vision while creating space for participants to explore practical, user and discipline-oriented decisions.
Table 1
Alignment of mentor interventions with the CASYE framework.
| CASYE COMPONENT | MENTORING FUNCTION | EXTRACT FROM MENTORING NOTES |
|---|---|---|
| Mentoring approach | Structured the start; clarified scope and expectations. | “Participants have clarity on the tool to develop: a chatbot for HR that can accomplish specific tasks, although they haven’t defined them yet.” |
| Idea development | Prompted function discussion of the OER. | “I asked several questions for reflection: What problem are you trying to solve with the chatbot? Who is going to use it, and how will it help them? Could it support recruitment or a specific area in the institutional context?” |
| Your potential | Recognized and reinforced use of prior experience. | “The strength of the group lies in their profiles; they have experience with generative AI tools, have worked together before, and developed OER.” |
| Your environment | Raised licensing concerns; prompted tool/platform choice. | “I suggested they check the compatibility of licenses. Does ChatGPT allow for the kind of open, educational reuse you’re planning?” They were considering ChatGPT. They decided to work with Poe instead.” |
| Your idea | Reinforced coherent structure and communicability of outputs. | “If you’re building additional materials make sure they follow the principles of OER. For the chatbot, I suggested they create a ‘catchy’ name related to human resources.” |
| Hard skills | Highlighted disciplinary contributions. | “Col5, who is a medical doctor and was a first-time participant, asked how he could contribute. The team suggested he contributed with content connecting HR to health.” |
| Soft skills | Supported inclusion, validated absent participants. | “They belong to the same university, are colleagues, and professors, which facilitates cohesion and collaboration. How can HR practices benefit from your expertise? Are other members joining the project?” |
| Team and ecosystem | Observed and supported team dynamics and leadership rotation. | “Col1 and Prom1 assumed leadership; tasks and decisions were distributed among the group”. |
| Access to finance | Not directly applicable in this case. | Not observed. All software and tools are used in the free trial version. |
| Legal forms | Reassured use of open licensing and attribution. | “Remember to get your open license from the Creative Commons website”. |
| Community engagement | Prompted alignment with 2030 Sustainable Development Goals (SDG) and social relevance. | “In your final pitch, focus on the chatbot demonstration rather than explanation. Link it to the SDG as it matters in the context of the lab. You will only have 10 minutes.” |
| Impact and sustainability | Framed product development with long-term use. | Not observed. All OER from the laboratory were placed in a repository by the coordinators. |
In Your environment, license verification and choice to work with Poe rather than ChatGPT illustrates how technical decisions become ethical acts. Then, mentoring in online collaborative environments supports not only academic and practical issues, but also social responsibility (Pollard & Kumar 2021).
Mentoring remained adaptive to the group’s characteristics. As observed in Hard and Soft skills, the validation of participants’ backgrounds and inquiry on absent members promotes the sense of belonging that is to be continuously fostered through social presence in virtual environments (Dulfer, Gowing & Mitchell 2024). Leadership roles were assumed rather than suggested by the mentor in Team and ecosystem, so it was not possible to observe its effect in terms of outcomes (Luo et al. 2022).
Moreover, the mentor’s emphasis on Creative Commons attribution (Legal forms) guided participants to consider the OER in the long-term (Community engagement and Impact and sustainability), expanding CASYE’s focus from entrepreneurship to ethical responsibility, corresponding with Marino et al., (2025). Access to finance was not observed as tools were used in the free version.
Although mentoring was implemented without a predefined model, in practice it aligned closely with the CASYE framework, confirming practical value. In the framework, Mentoring approach is a separate component of the cycle. As the mentor did not receive prior information about the participants, the mentoring approach had to be adapted as the project developed.
A critical thought is that the participants’ affiliation with the same institution, may have facilitated collaboration and their previous experience in developing OER may have overlapped with mentoring interventions, generating participants’ actions that could not be attributed to mentoring alone.
How do mentoring practices influence group dynamics, decision-making, and the collaborative development of a generative AI-based OER?
Lab-notebook entries in Table 2 show mentees’ actions and decisions for nearly every CASYE component. In Mentoring approach, once expectations were defined, roles and project deliverables were clarified. The group defined the chatbot HRM assistant functions, translating the mentor’s prompts into the design (Idea development). Disciplinary expertise shaped task distribution (Your potential). Mentees compared the chatbot’s ethical and licensing implications, reconsidering the mentor’s input on this (Your environment). All components of the OER were defined in Your idea (chatbot, manual, and tutorial video within a website).
Table 2
Alignment of mentees’ collaborative actions with the CASYE framework.
| CASYE COMPONENT | MENTORING FUNCTION | COLLABORATIVE ACTIONS | EXTRACT FROM LAB NOTEBOOK |
|---|---|---|---|
| Mentoring approach | Structured the start; clarified scope and expectations. | Assigned roles; clarified project scope and deliverables. | “The objective is to create a website that provides resources for orientation and learning to develop AI tools in HR.” |
| Idea development | Prompted function discussion of the OER. | Generated chatbot scenarios, defined content scope. | “The design chatbot in ChatGPT is to evaluate job positions and HR departments, and perform specific tasks related to industrial and organizational psychology”. |
| Your potential | Recognized and reinforced use of prior experience. | Mapped roles based on disciplinary expertise. | “Col1: theory and chatbot; Col11: theory for Topic 5; Col6, Col7, and Col5: onboarding content.” |
| Your environment | Raised licensing concerns; prompted tool/platform choice. | Compared Poe and ChatGPT, documented ethical/licensing decisions. | “Reviewed Poe because ChatGPT does not allow the type of free use we are looking for”. |
| Your idea | Reinforced coherent structure and communicability of outputs. | Outlined components (chatbot, manual, video), aligned them. | “We will use Wix to integrate the chatbot and the manual; we will also make a tutorial video.” |
| Hard skills | Highlighted disciplinary contributions. | Wrote manual, drafted chatbot scripts, built web site. | “This is what we need: Video procedure for chatbot development… Manual content: instructions, procedures, and chatbot development steps.” |
| Soft skills | Supported inclusion, validated absent participants. | Used WhatsApp for coordination. | “This is the recording for the meeting, if you couldn’t attend it today.” |
| Team and ecosystem | Observed and supported group dynamics and leadership rotation. | Rotated leadership, coordinated platform use, documented meetings. | “Assignments: webpage – Prom1; magazine – Prom2; theory and chatbot- Col1; chatbot- Col2 and Col4, theory- Col11.” |
| Access to finance | Not applicable in this case. | Not applicable. | Not observed. |
| Legal forms | Reassured use of open licensing and attribution. | Obtained a Creative Commons license. | The group got the Non-Commercial-Share-Alike license in the Creative Commons website. |
| Community engagement | Prompted alignment with SDG and social relevance. | Framed product as reusable for HR training and institutional use. | Resources included video, PDF and text processor technical documents, to promote access to educational resources across different contexts. |
| Impact and sustainability | Framed product development with long-term use. | Packaged product for accessibility. | The group published the final OER on the website. |
Entries for Hard and Soft skills tracked the technical contributions, collaboration and communication, even for absent participants. Team and ecosystem logs denoted leadership and shared responsibility. The technical documents and diversified resources added significant value to the AI-OER in Community engagement, Impact and sustainability, echoed the mentor’s suggestion regarding the final user and purpose.
Given the above, it is clear that mentoring practices aligned well with the group’s collaborative actions on how the group distributed responsibilities, and made design decisions. Rather than offering solutions and being directive, the mentor asked guiding questions and offered feedback across all the stages of the project leading participants to think critically about their options. This form of mentoring, as Baran (2020) states, scaffolds active participation and role negotiation among peers.
What are the characteristics of the generative AI-based OER developed by participants?
To provide an overview of the AI-OER developed, a general description of the website that hosts the chatbot is provided, followed by the results.
Four sections of the AI-OER were analyzed in this study: website, manual, videos and chatbots. The RH-REA website is organized around a horizontal menu with six main pages (Figure 2). Inicio is the home page. Presentación displays the context of the virtual laboratory, project and credits contributors. Manual links to a complete PDF guide to develop the chatbot, while Materiales concentrated HRM thematic guides in multiple formats for free download. Videos contains tutorials to create a chatbot for every area. Quick links refer to each one of the sections, making navigation easy and predictable. All pages show a Creative Commons license BY-NC-SA for users to retain, reuse, revise, remix and redistribute the resources. The site was developed in Spanish.
Figure 2
Home page AI-OER website.
Source: Recursos Humanos-REA https://jgaticap1.wixsite.com/rh-rea/copy-of-videos.
The interface of the AI chatbot is based on Poe AI (Figure 3). The central panel shows a conversation area to type messages or questions and see the bot’s replies. It was developed in Spanish but the multilingual chat function can switch to English or other languages.
Figure 3
Interface of AI- OER prototype.
Source: Recursos Humanos-REA. Bot RRHH 2024 https://poe.com/Bot-rrhh_2024.
The user’s manual “Organization innovation in 4.0 industry” (Figure 4) explains the development of the OER and how the idea originated. Also, the dynamics for virtual collaboration and a description for expert validation of contents are provided for the selected topics (Creation of my personalized bot, Design of the job description and position profile, Attraction, selection, and evaluation of human resources, Onboarding of human resources, and performance evaluation). This information is in Spanish only.
Figure 4
User’s manual for personalized AI chatbots.
Source: Adapted from Manual RH-REA https://jgaticap1.wixsite.com/rh-rea/acerca-de.
The Videos section contains links to tutorials for each of the HRM topics (Figure 5). They show step by step on how to configure each section of the bot. There are five tutorials in Spanish, linked to a YouTube channel.
Figure 5
Video tutorial to create a HRM chatbot.
Source: Recursos Humanos-REA. Laboratorio chatbot https://www.youtube.com/watch?v=s1qgugvgY-w.
The AI-OER was assessed against UNESCO (2022) licensing and accessibility principles for open educational resources. Table 3 summarizes the findings, outlining the strengths and limitations relative to such standards.
Table 3
Assessment of the AI chatbot across the OER principles.
| OER PRINCIPLE | EVALUATION | LEVEL OF ALIGNMENT | ANALYSIS |
|---|---|---|---|
| Open Licensing | The chatbot is covered under the CC BY-NC-SA license via the overall project documentation. | High | The chatbot can be freely used and adapted under a non-commercial license, which benefits organizations looking to train staff or support HR tasks. |
| Replicability | Instructions are provided to guide the development of a similar chatbot, but the original chatbot’s logic and scripting are not published. | Medium | Although the chatbot’s programming isn’t shared, practical instructions and examples help HRM teams replicate the experience with limited tech skills. |
| Adaptability | The chatbot scenarios are framed around HRM tasks, which can be adapted, though flexibility depends on users’ technical skill. | Medium | The design covers distinct HRM topics and could be tailored, but users cannot replicate or modify its architecture without recreating it in Poe independently. |
| Accessibility | It is accessible through the RH-REA Wix site without login or access barriers, available for immediate use. | High | Anyone can use the chatbot without logging in, making it easily accessible for HRM professionals or trainees. |
| Transparency | While intended use and design logic are explained in the manual, the inner workings are not exposed. | Medium | The manual explains what the chatbot does, but not how it works behind the scene, which might limit deep customization by users. |
| Usability | Users interact with the bot via predefined inputs; it is intuitive to use but offers limited interactivity and no adaptive responses. | Medium | The chatbot is simple and practical for typical HRM scenarios, though it doesn’t adjust its responses based on user behavior or learning progress. |
[i] Source: RH-REA website (https://jgaticap1.wixsite.com/rh-rea).
The chatbot aligns well with the foundational principles of open educational resources, particularly licensing and accessibility, granted through a CC BY-NC-SA license and hosted on a public website for educational and professional use. But as Norris, Swartz and Kuhlmeier (2023) noted, even when using a Creative Commons license, OER users often have difficulty in understanding its implications.
Replicability and transparency appeared partial limitations since, according to Bozkurt (2023), there is a grey area that creates conflict between creators conceding access and AI copyright. While users are provided with practical instructions on how to create a similar chatbot, the lack of access to the programming reduces the chatbot’s technical openness. A limitation that Bosch and Kruger (2024) argue may not affect immediate use but could restrict more advanced customization.
The adaptability of the chatbot is supported by a modular design such that it can be trained to address a range of HRM tasks. Still, deeper customization requires technical knowledge not covered by the current documentation. As Koenigstorfer et al. (2024) maintain, even users with some technical competence might have difficulties in developing AI educational tools because, without access to code or model logic, AI developments remain “black boxes”, and reuse is limited to interaction. In terms of usability, the chatbot is intuitive and clearly focused. However, the interaction model is still developing and not as dynamic or interactive as more advanced AI tools.
To complement the assessment of the AI-OER as a ‘package’, Table 4 presents a preliminary alignment check according to open access principles, using three levels defined for this exercise: Low (limited alignment), Medium (partial alignment), and High (full alignment). This rubric served as a basic verification, not a formally validated measurement tool.
Table 4
Assessment of the AI- OER package.
| COMPONENT | OPEN LICENSING | REPLICABILITY | ADAPTABILITY | ACCESSIBILITY | TRANSPARENCY |
|---|---|---|---|---|---|
| Website | Medium | Medium | Low | Medium | Medium |
| Manual | High | High | High | Low | High |
| Chatbot | Low | Low | Low | Medium | Medium |
| Video | Medium | Medium | Medium | Medium | Medium |
| Overall Package | Medium | Medium | Medium | Medium | High |
It can be said that the AI-based OER showed substantial alignment with UNESCO’s OER principles. Open licensing is used consistently in resources created by the group. Resources are replicable in multiple formats, and transparency is Medium acceptable. Adaptability and accessibility vary by medium. However, some sections of the website could be improved by adding introductory text rather than simply displaying the resources. The manual scored the highest because of its editable templates, while chatbots and videos depend on external platforms not controlled by the group. Overall, the AI-OER package scored Medium-High compliance, except for the chatbot, confirming Kooli’s (2023) findings that chatbot reuse is legally permitted only at surface level.
Conclusions
This study explored how mentoring interventions shaped the development of a generative AI-OER within a virtual social and educational innovation laboratory. By documenting both the process and the resulting products, the research offers insights into how digital mentoring, collaborative learning, and AI integration come together in the development of meaningful, open educational and professional chatbot tools.
In summary, regarding RQ 1, the findings revealed that mentoring was supportive and formative. The mentor’s role extended beyond content guidance; prompting ethical reflection, structuring collaboration, and helping participants frame their decisions in line with open educational values. Through intentional questioning and timely feedback, the mentor helped translate initial ideas into a product that was both coherent and contextually relevant, implementing fluid, non-directive and transformative coaching (Bird 2025; Maxwell, Hobson & Manning 2022).
As for RQ 2, collaborative actions were observable in how the roles emerged organically, the group adapting to challenges, and assuming responsibilities based on their strengths. Within this action-learning scenario, participants developed solutions to problems that they have to deal with in their lives (Ferguson et al. 2019). Decision-making moments, particularly around licensing, platform selection, and chatbot functionality were enhanced by the mentor’s ability to guide reflection and remind participants to maintain focus and a sense of direction.
In the case of RQ 3, while the AI-OER and its components do not fully meet all aspects for complete openness (Koenigstorfer et al. 2024), it was demonstrated to be an authentic contribution to open access and AI, as an accessible resource to assist with Human Resource Management tasks.
The study is significant in offering practical applications and providing guidelines for mentoring in virtual labs as well as the development of AI-OER. It is also worth noting that although this mentoring experience was positive, it confirms the need for more research on mentoring in higher education as observed by Tinoco-Giraldo, Torrecilla Sánchez and García-Peñalvo (2020) and limitations should be acknowledged.
Among other identified limitations are the scope of the study, involving one group within a two-week lab, which naturally restricts the amount of iteration, user testing, and evaluation of the post-implementation phase. The mentor had practised this role only once before, in a different laboratory at another institution with a different group; the development of an OER being the only common element between the two experiences. Additionally, the researcher-mentor’s dual role brings inherent biases even with actions in place to control it.
Despite this limited context, the mentoring practices aligned with those of the CASYE framework and seemed transferable across settings.
Nonetheless, the study confirmed mentoring as an active learning strategy in virtual, project-based contexts and illustrated how mentoring can support the development of an AI-OER, and educational content within an inclusive, interdisciplinary, and reflective learning process.
Implications and future research
Future research should explore longer-term implementations that allow for refinement, user testing, and the adoption of AI-OER into teaching and professional contexts. Comparative studies across institutions or disciplinary fields could provide a broader understanding of how mentoring functions under different contexts and within different frameworks. Additionally, further work is needed on how generative AI-generated content aligns with open licensing and reuse standards, as this technology is in constant evolution.
Ethics and Consent
The organizing committee authorized data collection within the duration of the laboratory, as long as ethical principles were observed. Participants provided electronic informed consent allowing the collaborative outputs to be used for academic research.
Acknowledgements
This research is supported by Mexico’s Secretariat of Science, Humanities, Technology and Innovation (SECIHTI) under the Postdoctoral Fellowships Program.
Competing Interests
The author has no competing interests to declare.