Building Teacher Agency with AI
Policy, Practice, and Pedagogy
Course Identity
Philosophy
The course's central thesis is that AI expands teacher capacity without replacing teacher judgment. The driving definition of “agency” is a three-part structure:
Know what you want → Know how to use the tools → Have the capacity to act
Teachers enter with pedagogical clarity. The course addresses the second and third parts. AI is framed as infrastructure for teacher agency — not a substitute for professional expertise.
A cohort model is used deliberately. Participants build relationships across all eight sessions and are expected to experiment, share failures, and learn from one another.
AIisinfrastructureforteacheragency.Notasubstituteforit.
Eight Sessions · Two Hours Each
AI Foundations
Establish the conceptual and relational foundation. Build accurate mental models of how LLMs work and create cohort identity.
- Definitions: AI, machine learning, large language models
- How generative AI predicts outputs (next-token prediction)
- Why hallucination happens and why it is rare but dangerous
- What LLMs are well-suited for vs. not
- Cohort norms: experiment without judgment, bring skepticism, respect varying AI familiarity
Deliverable: Structured written reflection (due before Session 2)
AI in Education Today
Ground Session 1 concepts in practice. Explore multiple tools, develop opinions about platform suitability, and begin thinking critically about how students use AI.
- Teacher use cases: lesson planning, rubric creation, differentiation, feedback, communication
- Platform landscape: foundational LLMs (Gemini, ChatGPT) vs. education-specific tools (MagicSchool AI)
- Student AI patterns: misuse (completing assignments, summarizing without reading) and productive use
- AI detection tools: how they work, what they cannot reliably determine, and appropriate use
- NotebookLM demonstration for productive student use cases
Deliverable: Structured reflection (due before Session 3)
Prompt Engineering & Tool Mastery
Develop the four-part prompt framework and master the full Gemini feature set. Participants leave with the start of a personal prompt library.
- Four-part prompt framework: Role – Task – Context – Format
- One-shot vs. iterative prompting; context window management
- Gemini features: Deep Research, Canvas, Learning Mode, file upload, image/video generation
- Memory and personalized instructions; models and settings
- Hands-on: live four-part framework demonstration on a group-chosen topic
Deliverable: First entry in personal prompt library + reflection (due before Session 4)
Risks, Limitations & NotebookLM
Address what can go wrong with AI — both mechanical reliability failures and behavioral over-reliance — and demonstrate NotebookLM as a mitigation tool.
- Hallucination: causes, how providers reduce it, where it remains most likely
- Citation errors and reference integrity failures: conditions and verification strategies
- Cognitive offloading: how outsourcing thinking reduces independent capability over time
- NotebookLM: source-grounded AI that confines outputs to user-uploaded documents
- Features: source upload, study guide, FAQ, podcast-style audio, sharing
Deliverable: Personal AI safety checklist + reflection (due before Session 5)
AI Policy, Privacy Law & Guidelines
Situate practical AI skills within the policy landscape teachers operate in. Participants leave knowing legal obligations, professional standards, and cohort-defined use principles.
- District level: county/district AI policy and acceptable use guidelines (customized per cohort)
- State level: Maryland MSDE guidance; Blueprint for Maryland's Future (2021)
- FERPA: what constitutes student data; identifiability rules; what cannot enter AI tools
- COPPA: requirements for tools used with students under 13
- UNESCO and ISTE AI frameworks for professional practice
Deliverable: Cohort AI use guidelines (group) + individual reflection (due before Session 6)
AI-Assisted Grading & Academic Integrity
Build a clear framework for where AI can and cannot be trusted in grading, with hands-on comparison of AI-only, AI-assisted, and human-only grading.
- Grading criteria framework: Quantitative (AI reliable) / True-False (AI reliable with verification) / Qualitative (teacher judgment required)
- Hands-on: grade the same anonymized sample assignment in three conditions and compare
- AI detection with ZeroGPT: how it works and its critical limitations (strong writers trigger false positives)
- In-person authorship verification strategies for essays, research, math, and any assignment
Deliverable: Reflection (due before Session 7)
Student AI Use, Policy & AI-Resilient Assignments
Shift focus to the student side. Frame AI misuse as a motivational symptom, not a compliance problem. Participants leave with a working classroom AI use policy.
- Motivation framework: extrinsic (grade-driven) vs. intrinsic (meaning-driven) — why AI breaks grade-as-proxy-for-learning
- The 3X Framework: Exploration, Expression, Extension — assignment design that builds genuine engagement
- Why blanket AI bans are not effective and what works instead
- Classroom AI use policy components: permitted uses, prohibited uses, disclosure, consequences, student rights
Deliverable: Draft classroom AI use policy + reflection + Belbin team inventory (due before Session 8)
Final Project: Student AI Literacy Lesson
Culminating session. Participants synthesize all eight sessions into a classroom-ready AI literacy lesson for their own students.
- Build phase (60 min): groups of 4 (formed via Belbin inventory) construct a complete 1-hour AI literacy lesson
- Critique Round 1 (20 min): groups exchange lessons and provide written feedback
- Critique Round 2 (20 min): groups exchange with a third group for a second critique round
- Cohort synthesis discussion: reflection on overall shift in perspective across the arc of the course
Deliverable: Final AI literacy lesson (revised post-session) + take-home final reflection (due within one week)
Four Frameworks Used Throughout the Course
Teacher Agency
Know what you want → Know how to use the tools → Have the capacity to act. Teachers enter with pedagogical clarity; the course addresses the second and third parts.
4-Part Prompt Framework
Role – Task – Context – Format. The Context component carries the most leverage; most teachers underinvest here. Introduced in Session 3 and used throughout.
3X Framework
Exploration – Expression – Extension. An assignment design model for building genuine student engagement that AI cannot shortcut.
Grading Criteria Tiers
Quantitative / True-False / Qualitative. Determines where AI can assist in grading and where teacher judgment is non-negotiable.
What Participants Build Across the Course
Written reflection after every session
Personal prompt library (first entry; added to across the course)
Personal AI safety checklist
Cohort AI use guidelines (collaborative group document)
Draft classroom AI use policy for your subject area
Final student AI literacy lesson (peer-reviewed and revised)
Take-home final reflection (due within one week of Session 8)
CPD Credit
Once all requirements are met, Helicon AI Institute issues a signed MSDE CPD credit sheet confirming completion of Building Teacher Agency with AI (Approval #26-68-06) for 1 CPD credit (18 hours).
Requirements include attendance across all 8 sessions, submission of all session deliverables, and completion of the final project.
Enroll or Learn MoreGet in Touch
Questions about the course, enrollment, or bringing a cohort to your district? Contact us.