2026 AI Educator Summer Institute

*This is a cohort session. Following one morning session on Monday, July 13, this group will meet every afternoon to continue its exploration and learning as a cohort.*

AI models—such as Google Gemini, Perplexity, ChatGPT, Claude, etc. — offer teachers the opportunity to quickly review and research standard mathematics topics taught in high school and college courses.  In this session, participants will have the opportunity to compare classical solution methods with AI-generated solutions, and in the process educators will gain a deeper understanding of the relative strengths of each of these approaches.  The session will illustrate how AI resources can be useful as collaborative partners, offering inspiration for new methodologies, suggesting generalizations, and identifying similar problems to enrich the learning process.

*This is a cohort session. All participants who attend the morning session are invited to attend an accompanying afternoon expansion pack session each afternoon this week to apply your morning learning to the Special Education classroom.*

Special educators carry an extraordinary load. Differentiated instruction, IEP documentation, progress monitoring, parent communication, and the daily work of meeting every student exactly where they are. AI tools cannot replace the human expertise at the heart of that work. But they can extend it.

This session builds a practical, critical framework for using AI in inclusive classrooms and special education contexts. We explore what AI does well, including generating differentiated materials, audio versions of text, scaffolded activities, and adapted assessments. We also examine where it fails. From hallucination to awareness of bias and privacy risks that matter, especially when working with vulnerable students.

Participants work hands-on with Google Gemini and NotebookLM to adapt reading levels, generate modified materials, and explore AI-assisted documentation. Data privacy, FERPA, and keeping human judgment at the center of every decision are woven throughout, not added at the end.

*This is a cohort session. All participants who attend the morning session are invited to attend an accompanying afternoon expansion pack session each afternoon this week to apply your morning learning to the Special Education classroom.*

The 3-hour afternoon workshop gives participants dedicated creation time to apply the morning's framework to real materials from their own classrooms. It is structured as a creation studio rather than a continuation of the lecture.

Hour 1 (1:00–2:00) — Identify and Design: Participants select one real instructional challenge. A lesson that needs differentiation, a text that needs to be adapted for a lower reading level, an assessment that needs scaffolding, or a documentation task they want to streamline. Small groups form around shared challenges. Special educators learn as much from each other as from any presenter. Introduction to the creation workflow: original material → AI prompt → draft → critical review → revision.

Break (2:00–2:15)

Hour 2 (2:15–3:15) — Create: Hands-on creation using Gemini and NotebookLM. Facilitator circulates with targeted coaching. Structured peer feedback in pairs: Does this serve the student? What did the AI miss that a human expert would catch? What needs to be fixed before this goes near a classroom?

Hour 3 (3:15–4:00) — Share, Critique, and Plan: Small group shares. Not just "look what I made" but "here's what the AI got wrong and how I fixed it." Whole-group reflection on AI as a tool for inclusive practice versus a shortcut that bypasses expertise. Individual action planning to identify one specific thing I will try with students in fall 2026.

Participants leave with a real, usable adapted resource and a realistic, critical framework for continuing independently.

*This is a cohort session. All participants who attend the morning session are invited to attend an accompanying afternoon expansion pack session each afternoon this week to apply your morning learning to the Special Education classroom.*

The 3-hour afternoon workshop gives participants dedicated creation time to apply the morning's framework to real materials from their own classrooms. It is structured as a creation studio rather than a continuation of the lecture.

Hour 1 (1:00–2:00) — Identify and Design: Participants select one real instructional challenge. A lesson that needs differentiation, a text that needs to be adapted for a lower reading level, an assessment that needs scaffolding, or a documentation task they want to streamline. Small groups form around shared challenges. Special educators learn as much from each other as from any presenter. Introduction to the creation workflow: original material → AI prompt → draft → critical review → revision.

Break (2:00–2:15)

Hour 2 (2:15–3:15) — Create: Hands-on creation using Gemini and NotebookLM. Facilitator circulates with targeted coaching. Structured peer feedback in pairs: Does this serve the student? What did the AI miss that a human expert would catch? What needs to be fixed before this goes near a classroom?

Hour 3 (3:15–4:00) — Share, Critique, and Plan: Small group shares. Not just "look what I made" but "here's what the AI got wrong and how I fixed it." Whole-group reflection on AI as a tool for inclusive practice versus a shortcut that bypasses expertise. Individual action planning to identify one specific thing I will try with students in fall 2026.

Participants leave with a real, usable adapted resource and a realistic, critical framework for continuing independently.

This session will engage participants in the development of a cross-disciplinary data science course aimed at students and teachers with no prior coding background. It introduces the fundamentals of Data Science across subject areas with an introduction of Python and  with Jupyter notebooks. This session will challenge participants to imagine an introductory CS course that cuts across all subjects and can be taught by someone from any department with a little motivation. All backgrounds, subject areas, and skill-levels welcome.

This session will engage participants in the development of a cross-disciplinary data science course aimed at students and teachers with no prior coding background. It introduces the fundamentals of Data Science across subject areas with an introduction of Python and  with Jupyter notebooks. This session will challenge participants to imagine an introductory CS course that cuts across all subjects and can be taught by someone from any department with a little motivation. All backgrounds, subject areas, and skill-levels welcome.

*This is a cohort session. All participants who attend the morning session are invited to attend an accompanying afternoon expansion pack session each afternoon this week to apply your morning learning to the Special Education classroom.*

The 3-hour afternoon workshop gives participants dedicated creation time to apply the morning's framework to real materials from their own classrooms. It is structured as a creation studio rather than a continuation of the lecture.

Hour 1 (1:00–2:00) — Identify and Design: Participants select one real instructional challenge. A lesson that needs differentiation, a text that needs to be adapted for a lower reading level, an assessment that needs scaffolding, or a documentation task they want to streamline. Small groups form around shared challenges. Special educators learn as much from each other as from any presenter. Introduction to the creation workflow: original material → AI prompt → draft → critical review → revision.

Break (2:00–2:15)

Hour 2 (2:15–3:15) — Create: Hands-on creation using Gemini and NotebookLM. Facilitator circulates with targeted coaching. Structured peer feedback in pairs: Does this serve the student? What did the AI miss that a human expert would catch? What needs to be fixed before this goes near a classroom?

Hour 3 (3:15–4:00) — Share, Critique, and Plan: Small group shares. Not just "look what I made" but "here's what the AI got wrong and how I fixed it." Whole-group reflection on AI as a tool for inclusive practice versus a shortcut that bypasses expertise. Individual action planning to identify one specific thing I will try with students in fall 2026.

Participants leave with a real, usable adapted resource and a realistic, critical framework for continuing independently.

Now that participants understand the goal and exciting potential of this course, we will start exploring the design and content of the course. In the first part of the workshop we will do a group data science activity to get used to Jupyter Notebooks and Python. In the second part of the workshop attendees will be given time to explore resources and build their own lesson in a chosen disciplinary area with the aim of building a library of activities across disciplines that could be used by this introductory course for teachers across the state. 

*This is a cohort session. All participants who attend the morning session are invited to attend an accompanying afternoon expansion pack session each afternoon this week to apply your morning learning to the Special Education classroom.*

The 3-hour afternoon workshop gives participants dedicated creation time to apply the morning's framework to real materials from their own classrooms. It is structured as a creation studio rather than a continuation of the lecture.

Hour 1 (1:00–2:00) — Identify and Design: Participants select one real instructional challenge. A lesson that needs differentiation, a text that needs to be adapted for a lower reading level, an assessment that needs scaffolding, or a documentation task they want to streamline. Small groups form around shared challenges. Special educators learn as much from each other as from any presenter. Introduction to the creation workflow: original material → AI prompt → draft → critical review → revision.

Break (2:00–2:15)

Hour 2 (2:15–3:15) — Create: Hands-on creation using Gemini and NotebookLM. Facilitator circulates with targeted coaching. Structured peer feedback in pairs: Does this serve the student? What did the AI miss that a human expert would catch? What needs to be fixed before this goes near a classroom?

Hour 3 (3:15–4:00) — Share, Critique, and Plan: Small group shares. Not just "look what I made" but "here's what the AI got wrong and how I fixed it." Whole-group reflection on AI as a tool for inclusive practice versus a shortcut that bypasses expertise. Individual action planning to identify one specific thing I will try with students in fall 2026.

Participants leave with a real, usable adapted resource and a realistic, critical framework for continuing independently.

Now that participants understand the goal and exciting potential of this course, we will start exploring the design and content of the course. In the first part of the workshop we will do a group data science activity to get used to Jupyter Notebooks and Python. In the second part of the workshop attendees will be given time to explore resources and build their own lesson in a chosen disciplinary area with the aim of building a library of activities across disciplines that could be used by this introductory course for teachers across the state. 

This session equips educators to address key limitations of AI tools, including bias and hallucination, often caused by reliance on uncontrolled or unreliable sources. It provides participants with the knowledge and hands-on experience needed to guide AI toward generating insights from curated, reliable, and controlled resources.    Participants will begin by comparing open- and closed-source AI tools, developing a clear understanding of their strengths, limitations, and appropriate use cases. They will then explore two project examples created with a closed-source platform such as Google NotebookLM, examining how these models can be adapted for technology and engineering classroom contexts. Next, participants will develop their own project and resource library on a topic of personal or curricular relevance, emphasizing organization, content curation, and instructional design. By the end of the session, educators will leave with a project in progress and a practical framework for selecting and applying AI tools in teaching.

*This is a cohort session. All participants who attend the morning session are invited to attend an accompanying afternoon expansion pack session each afternoon this week to apply your morning learning to the Special Education classroom.*

The 3-hour afternoon workshop gives participants dedicated creation time to apply the morning's framework to real materials from their own classrooms. It is structured as a creation studio rather than a continuation of the lecture.

Hour 1 (1:00–2:00) — Identify and Design: Participants select one real instructional challenge. A lesson that needs differentiation, a text that needs to be adapted for a lower reading level, an assessment that needs scaffolding, or a documentation task they want to streamline. Small groups form around shared challenges. Special educators learn as much from each other as from any presenter. Introduction to the creation workflow: original material → AI prompt → draft → critical review → revision.

Break (2:00–2:15)

Hour 2 (2:15–3:15) — Create: Hands-on creation using Gemini and NotebookLM. Facilitator circulates with targeted coaching. Structured peer feedback in pairs: Does this serve the student? What did the AI miss that a human expert would catch? What needs to be fixed before this goes near a classroom?

Hour 3 (3:15–4:00) — Share, Critique, and Plan: Small group shares. Not just "look what I made" but "here's what the AI got wrong and how I fixed it." Whole-group reflection on AI as a tool for inclusive practice versus a shortcut that bypasses expertise. Individual action planning to identify one specific thing I will try with students in fall 2026.

Participants leave with a real, usable adapted resource and a realistic, critical framework for continuing independently.

*This is a cohort session. All participants who attend the morning session are invited to attend an accompanying afternoon expansion pack session each afternoon this week to apply your morning learning to the Special Education classroom.*

The 3-hour afternoon workshop gives participants dedicated creation time to apply the morning's framework to real materials from their own classrooms. It is structured as a creation studio rather than a continuation of the lecture.

Hour 1 (1:00–2:00) — Identify and Design: Participants select one real instructional challenge. A lesson that needs differentiation, a text that needs to be adapted for a lower reading level, an assessment that needs scaffolding, or a documentation task they want to streamline. Small groups form around shared challenges. Special educators learn as much from each other as from any presenter. Introduction to the creation workflow: original material → AI prompt → draft → critical review → revision.

Break (2:00–2:15)

Hour 2 (2:15–3:15) — Create: Hands-on creation using Gemini and NotebookLM. Facilitator circulates with targeted coaching. Structured peer feedback in pairs: Does this serve the student? What did the AI miss that a human expert would catch? What needs to be fixed before this goes near a classroom?

Hour 3 (3:15–4:00) — Share, Critique, and Plan: Small group shares. Not just "look what I made" but "here's what the AI got wrong and how I fixed it." Whole-group reflection on AI as a tool for inclusive practice versus a shortcut that bypasses expertise. Individual action planning to identify one specific thing I will try with students in fall 2026.

Participants leave with a real, usable adapted resource and a realistic, critical framework for continuing independently.

By the end of this session, participants will be able to:

1. Identify at least two ways AI systems reflect bias and explain why this matters for classroom learning and student agency.
2. Recognize human cognitive biases — including confirmation bias and authority bias — that shape how students and teachers interact with AI outputs.
3. Apply a framework for evaluating student AI use that centers critical thinking and responsible judgment, not detection or compliance.
4. Redesign or adapt at least one existing task to support transparent, integrity-centered AI integration using a prompt rubric and transcript as the assessable artifact.