AI Study Mode

For those of you who will begin your academic term soon, welcome back to teaching and to this scholarly SoTL blog! For those who are on the quarter system, it seems you have another month, but chances are you are already preparing for classes. 

Over the summer, like many of you, I have been playing intensely with how AI can be integrated into our teaching and learning in a meaningful way. So, this week I would like to share a recent development from OpenAI called Study Mode. Study mode is a built-in ChatGPT mode that turns the assistant into a tutor. Instead of just giving answers, it guides you step-by-step with Socratic questions, scaffolded explanations, and formative assessments that adapt to your goals and level (using memory from the conversation). I created an example using one of the assignments in a class that I typically teach, in case you would like to view how this could look.

Study mode represents a deliberate move toward aligning AI with evidence-based learning science. By using scaffolded, interactive guidance rather than direct answer delivery, study mode fosters active engagement, metacognition, and self-regulated learning.

AI tools have often been criticized for enabling passive “answer retrieval” rather than fostering deep learning. Study mode applies principles from How People Learn (Bransford, Brown, & Cocking, 2000), the ICAP engagement framework (Chi & Wylie, 2014), and cognitive load theory (Sweller, Ayres, & Kalyuga, 2011) to create a more purposeful, student-centered interaction using a stepwise scaffold approach.

Step-by-Step Scaffold

1. Establish Baseline Understanding.

2. Elicit Prior Knowledge

3. Expand the Solution Space

4. Refine Through Critical Inquiry

5. Synthesize a Combined Approach

6. Integrate Applied Consideration

Implications for Teaching and Learning with AI

Study mode illustrates how AI can operationalize decades of learning science research:

• Supports constructivist learning by building on the student’s prior knowledge.

• Encourages cognitive apprenticeship through guided practice in expert reasoning.

• Fosters self-regulation by prompting learners to make decisions and justify them.

• Bridges theory and practice by requiring learners to apply domain concepts to authentic, complex scenarios.

Study mode offers an instructional design pattern that mirrors the best practices of human tutoring: diagnosing needs, scaffolding knowledge, eliciting active engagement, and gradually handing over cognitive control to the learner. When paired with sound pedagogy, AI can support not just knowledge acquisition but the higher-order reasoning, adaptability, and reflective judgment that education strives to cultivate.

Also for those of you who are playing with creating Custom GPTs, I have developed a few and would like to share in case they are of interest (note, you do need a free OpenAI account to access these)::

1. QualiText Analyst analyzes qualitative data to provide keyword extraction, sentiment analysis, thematic analysis, and clustering.

2. Teaching Statement Guide helps faculty convey core ideas about being an effective teacher. Develops ideas with examples of how learners will achieve outcomes.

3. Research Agenda Guide is a summary of research accomplishments, current work, and future direction and potential of scholarship.

4. Effective Teaching Evidence Guide uses Wiggins & McTighe backward design. Evidence includes peer/self-evaluation, videos, teaching research, and portfolio.

5. SoTL Writing Mentor supports faculty who are interested in collecting data on their teaching effectiveness to prepare scholarly manuscripts. 

6. Course Design Studio Coach guides faculty through the course design process.

7. Course AI Policy Statement Guide is designed to help faculty craft clear, values-aligned AI policy statements for their courses.

8. GPT as a Teaching Assistant is designed to enhance engagement and draws on foundational materials such as a teaching philosophy, research agenda, syllabus, assignments, assessment rubrics, and scholarly publications.

9. Assessment and AI Coach supports AI-integrated assessment design, academic integrity, and creativity in education.

10. Supplemental Instruction, SI Companion provides an effective teaching research approach to SIs, so they can support students at a deeper level. 

And finally, for those of you who are interested in using AI to create dynamic electronic learning objects (eLOs), you may want to look into Google's NotebookLM. This tool is part of the Google suite and allows three free podcast/screencasts per day (if you have an institutional account, you can create more). I am in the process of creating these for my scholarly papers. Here is an example of each tool:

• Hill, C., & Hargis, J. (Sep 2024). An ethics module on academic integrity and generative AI. New Directions for Teaching and Learning, 1–10. 

  • NotebookLM Podcast Audio

  • NotebookLM Screencast Video

References

Bransford, J. D., Brown, A. L., & Cocking, R. R. (Eds.). (2000). How people learn: Brain, mind, experience, and school(Expanded ed.). Washington, DC: National Academies Press.

Chi, M. T. H., & Wylie, R. (2014). The ICAP framework: Linking cognitive engagement to active learning outcomes. Educational Psychologist, 49(4), 219–243. https://doi.org/10.1080/00461520.2014.965823

Sweller, J., Ayres, P., & Kalyuga, S. (2011). Cognitive load theory. Springer Science & Business Media.

Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Harvard University Press.