Self-Determination & Generative AI: High School Student-Led Mixed-Age AI Learning

Framework

In 2022 ChatGPT 3.5 Generative Artificial Intelligence (GAI) took many schools by surprise, creating research opportunities (Wong & Looi 2024; Baidoo-Anu & Ansah 2023). Future work will benefit from student experience effectively leveraging GAI and managing collaborative human teams. Here, we show an approach to teaching and learning with GAI in a setting authentic to developing future-of-work skills. High school-developed questions are part of a larger research project, representing student-identified gaps in understanding among their peers and educators. Learning to lead in complex mixed-age settings, high school students contribute analysis of their lived experiences, supporting iterative curriculum development and educator training with GAI within a larger research initiative that received IRB approval.

Self-determination theory and intrinsic motivation (Ryan & Deci, 2020) meet Fair et al. (2005) paralleled roles of expert/novice and roles of caregiver/child for "intellectual tools and skills in the context of joint involvement" (p. 350). Kellett (2010) acknowledged, “adults have greater knowledge than children in many areas of life but with regard to childhood itself- in the sense of what it is like to be a child - it is children who have the expert knowledge" (p. 197). Generative Artificial Intelligence is socially-culturally situated (Cole, 1990).

Methods

Action research is a "way to improve practice" (Glesne, 2006, p. 17) as well as "finding a solution to a local problem in a local setting" and is conducted to obtain information and knowledge pertaining to a complex practical problem about which not much is known, and no heuristics exist (Mills, 2003), while grounded theory assisted in identifying emerging patterns (Charmaz, 1995). Bridging research and practice with GAI must include student-informed perspective through PAR (Langhout & Thomas 2010; Phillips & Carr 2006; Pultorak et. al. 2006). The combination of these became part of both methodology and analysis.

Student-developed research questions:

RQ: What does self-determination theory look like in high school students at our school as they develop human skills with GAI support while leading mixed-age project-based learning groups?

SQ1: What GAI prompts do older students generate to support their leadership skills?

SQ2: In what ways do older students identify and revise their prompts to facilitate  achieving goals, including guiding less-experienced learners in mixed-age groups?

SQ3: When older students prompt GAI for learning process guidance, what thought  process changes are observable in older students as they persist to achieve project goals despite challenges?

Data Sources
Artifacts including student-created video, images, notes, class projects, reports, Teams audio call recorded analysis of discussions, field notes, and prompts and Generative Artificial Intelligence content and screen shots with Magic School AI, Copilot, and ChatGPT, as well as GAI tools within Canva, and other presentation sources.

Key features
Digital interactive student artifact collections illustrate and are digitally accessible to viewers. Examples of interactive artifacts as data sources are available here [please do not navigate to link prior to completion of blind peer review]
https://wakelet.com/wake/VRWFMjSrZwW_j0P7MH7Rm. Early analysis revealed actionable approaches where GAI may complement self-determination theory, facilitating intrinsic motivation development in older students while leading and meeting goals with mixed-age groups.

Participant Selection
The school established a school-university research partnership in 2024 and industry partnership for over a decade. It is a small school with origins back to the 1960s in one of the most diverse counties in Washington State. The school serves students from Kindergarten-Prep through 10th grade, offering a college prep program for 9th and 10th graders looking to enter college early through a state-funded dual enrollment program. The school is known for a diverse student body, representing over 27 ethnic backgrounds and 14 faith traditions. The school's mission includes preparing students to contribute solutions to the challenges of our global community, while learning from and respecting multiple perspectives, including those who think differently from them and have a different view the world. Participants included those who volunteered to share their work and communicate their ideas across adults (teachers and university professors), and students as peers and less-experienced mentees. Older students were part of iterative design and prior professional development conferences and had previously contributed ideas and policy for the school's GAI policies. This iteration of the design included older students collaborating with educators to refine their questions as research questions, identify gaps they wished to address, and iterate on design co-developed by a core group of educators, then deployed to classroom teachers (including those with little to no experience with generative artificial intelligence). Students collected data, debriefed, and curated content to communicate their findings to adult professional audiences at ASU+GSV AI Show. They returned with feedback from adults and iterated on their design another time before the end of the 2024-2025 school year where data collection was completed.

Results

High School students successfully identified gaps that needed to be addressed before implementing generative artificial intelligence in professional development for educators as an outcome of pursuing their collaboratively developed research questions:

RQ: What does self-determination theory look like in high school students at our school as they develop human skills with GAI support while leading mixed-age project-based learning groups?

SQ1: What GAI prompts do older students generate to support their leadership skills?

SQ2: In what ways do older students identify and revise their prompts to facilitate  achieving goals, including guiding less-experienced learners in mixed-age groups?

SQ3: When older students prompt GAI for learning process guidance, what thought  process changes are observable in older students as they persist to achieve project goals despite challenges?

Through several focused domains, high school students explored applications of GAI in an intersection of coding and dance, STEM stomp rockets, and social studies focus on Black History Month. They selected these domains of interdisciplinary learning to span Computer Science, performing arts, social studies, and STEM to explore challenges and benefits for using AI in each and if each domain required a different response, prompt-development, or application. They identified variations and pain points in their own development of content, and then a second round of challenges when implementing the lesson plans either entirely without the support of GAI, fully developed by GAI or a hybrid. With hybrid models, high school students were able to refine which aspects of the lesson plans were most effective on the educator side (their teaching side) and which benefitted the students most. Their products were presented at ASU+GSV AI Show, where they obtained feedback and a final iteration was completed. From there, our research team re-designed professional development and involved teachers in co-design of AI Studios beginning fall 2025 based on the findings of the high school students and modeling in the classrooms.

One of the most notable findings the high school researchers determined was that they would not choose to use GAI to replace lesson design, but facilitate their organization, aligning standards, age-grade alignment ideas, but they found assessment and creative approaches to personalize learning was best served by their knowledge as the lesson co-designers and knowing their "students" in a way GAI did not. They also found that depending on their prompt-engineering, lesson plans entirely generated by AI returned relatively generic results or drawing from the most common information publicly available. This did not serve them well in developing lessons that challenged creative approaches and ISTE standards, or supporting students to produce unique outcomes, which was the goal as prior iterations and study by this high school cohort delved into irreplaceable content creation unique to students (vs students allowing GAI to produce content for them to turn in as a completed assignment). Because of their prior learning and defining "cheating" vs assignments that support practicing and developing specific skills, these students saw nuance in lesson plan design that they felt was important to maintain vs turn over to GAI, and wanted adults to be aware of those considerations.

In terms of motivation, they found that their personal intrinsic motivation to persist in difficult situations increased when they felt the combination of relatedness, autonomy and competence. They were able to articulate similar patterns in the younger students they taught. When one or more of those innate psychological needs were not met, there was a decrease in authentic intrinsic motivation to persist in challenging tasks (regardless of the domain). While not their primary aim, the high school students also disclosed to the lead researcher what they observed in the teachers they worked with to integrate their lesson plans and their observations on apparent types of motivation from teachers (whether intrinsic - wanting to learn and apply something new, seeing value from high school students experiences, or extrinsic- "because they had to"). A final analysis will be completed prior to ISTE 2026.

Importance

To our knowledge, this IRB contribution is first of its kind at the intersection of GAI, human-computer interaction, motivation theory, student lived experiences developing real-world future-of-work skills, defined from their perspective, for the purpose of influencing iterative curriculum development and implementation (Arlington 2001; Ross 2006; Ryan & Deci 2020) as part of standard educational procedure. This contribution is also unique in that it is not only presented in written format, but digital artifacts that allow for replication in other classrooms bringing theory and research into practice.

References

Author (2011; 2016; 2018; 2022; 2024)

Akgun, S., & Greenhow, C. (2021). Artificial intelligence in education: Addressing ethical challenges in K-12 settings. AI and Ethics, 2(3).

Arlington, J. W. (2001). The future of theory-of-mind research: understanding motivational states, the role of language, and real-world consequences. Child Development, 72(3), 411-568.

Baidoo-Anu, D., & Leticia Owusu Ansah. (2023). Education in the Era of Generative Artificial Intelligence (AI): Understanding the Potential Benefits of ChatGPT in Promoting Teaching and Learning. Journal of AI, 7(1), 52–62. https://doi.org/10.61969/jai.1337500

Charmaz, K. (1995). Grounded theory. In H. Smith & V. Langenhove (Eds.) Rethinking Methods in Psychology (pp. 27-49). Newbury Park, CA: Sage Publications Ltd.

Counts, G. S. (1978). Dare the school build a new social order?. Carbondale: Southern Illinois University Press.

Cotton, D. & Gretsy, K. (2006). Reflecting on the think-aloud method for evaluating e learning. British Journal of Educational Technology, 37(1), 45-54.

Glesne, C. (2006). Becoming qualitative researchers: An introduction (3rd Edition). Boston, MA: Pearson.

Hunleth, J. (2011) Beyond on or with: Questioning power dynamics and knowledge production in ‘child oriented’ research methodology. Childhood, 18(1), 81-93, doi: 10.1177/0907568210371234

Human-Centered Artificial Intelligence in Schools. (n.d.). Ospi.k12.Wa.us. https://ospi.k12.wa.us/student-success/resources-subject-area/human-centered-artificial-intelligence-schools

Langhout, R. D., & Thomas, E. (2010). Imagining participatory action research in collaboration with children: An introduction. American Journal of Community Psychology. eScholarship, University of California. 46, 60-66. Doi: 10.1007/s10464-010-9321-1

Karcher, M.J. (2008). The cross-age mentoring program: A developmental intervention for promoting students’ connectedness across grade levels. Professional School Counseling, 12(2), 137-143.

Kellett, M. (2005). How to develop children as researchers: A step by step guide to teaching the research process. Thousand Oaks, CA. SAGE Publications Inc.

Khan, S. (2024). Brave New Words. Penguin.

Kidd, S., & Kral, M. J. (2005). Practicing participatory action research. Journal of Counseling Psychology, 52(2), 187-195. doi: 10.1037/0022-0167.52.2.187

Klar, S., Krupnikov, Y., Ryan, J. B., Searles, K., & Shmargad, Y. (2020). Using social media to promote academic research: Identifying the benefits of twitter for sharing academic work. PloS one, 15(4), e0229446. https://doi.org/10.1371/journal.pone.0229446

Meriam, S. B. (1998). Qualitative research and case study applications in education. (2nd Edition). San Francisco, CA: Jossey-Bass.

Microsoft Learn. (2024). Education Transformation Framework - Training. Retrieved March 19, 2024, from https://learn.microsoft.com/en-us/training/courses/education-transformation-framework

Mishra, P., Oster, N., & Henriksen, D. (2024). Generative AI, Teacher Knowledge and Educational Research: Bridging Short- and Long-Term Perspectives. TechTrends. https://doi.org/10.1007/s11528-024-00938-1

Noonoo, S. (2023). Setting Ground Rules Around Original Writing and ChatGPT. Edutopia. Retrieved March 18, 2024, from https://www.edutopia.org/article/ground-rules-plagiarism-original-writing-chatgpt

March, J. (1999). Yo sé quien soy. In Kenneth Sirotnik and Roger Soder, (Eds.), The beat of a different drummer: essays on educational renewal in honor of John I. Goodlad. Peter Lang; New York. Pp. 273-280.

Wooltorton, S., Collard, L., Horwitz, P., Poelina, A., & Palmer, D. (2020). Sharing a place-based indigenous methodology and learnings. Environmental Education Research, 26(7), 917–934. https://doi.org/10.1080/13504622.2020.1773407

Palinkas, L. A., Horwitz, S. M., Green, C. A., Wisdom, J. P., Duan, N., & Hoagwood, K. (2015). Purposeful

Sampling for Qualitative Data Collection and Analysis in Mixed Method Implementation Research. Administration and Policy in Mental Health and Mental Health Services Research, 42(5), 533–544. PubMed Central. https://doi.org/10.1007/s10488-013-0528-y

Phillips, D. K., & Carr, K. (2006). Becoming a teacher through action research; Process, context, and self-study. New York: Routledge.

Pultorak, E., McCarthy, J., & Young, M.W. (2006). School and university partnerships and the role of teacher as researcher. Action in Teacher Education, 28, 1, 99-109.

Ross, H. (2006). Theory of Mind Goes Social. Human Development 49(6), 375-379.

Ryan, R. M., & Deci, E. L. (2020). Intrinsic and extrinsic motivation from a self-determination theory perspective: Definitions, theory, practices, and future directions. Contemporary Educational Psychology, 61, Article 101860. https://doi.org/10.1016/j.cedpsych.2020.101860

Shaheen, M., Pradhan, S., & Ranajee, R. (2019). “Sampling in Qualitative Research.” Advances in Business Information Systems and Analytics, 25–51. https://doi.org/10.4018/978-1-5225-5366-3.ch002

Steele, C. M. (1997). How stereotypes shape intellectual identity and performance. American Psychologist, 52(6), 613-629.

Tillman, L. C. (2006). Researching and writing from an African-American perspective: Reflective notes on three research studies. International Journal of Qualitative Studies in Education, 19(3), 265-287.

United Nations. (2023). Education | Department of Economic and Social Affairs. Sdgs.un.org. https://sdgs.un.org/topics/education

Wong, L.-H., & Chee-Kit Looi. (2024). Advancing the generative AI in education research agenda: Insights

from the Asia-Pacific region. Asia Pacific Journal of Education, 44(1), 1–7. https://doi.org/10.1080/02188791.2024.2315704

Yu, H., & Guo, Y. (2023). Generative artificial intelligence empowers educational reform: current status, issues, and prospects. Frontiers in Education, 8. https://doi.org/10.3389/feduc.2023.1183162