About 50 teachers gathered in New York on March 18 to learn “agentic” AI tools—autonomous systems that can perform complex, multi-step tasks with reasoning capabilities. The training, part of the National Academy for AI Instruction, represents a shift from surface-level AI use to sophisticated instructional applications.

• Agentic tools: AI systems that reason and execute multi-step tasks
• Lesson stress-testing: Tools that identify content gaps and confusing wording
• Professional judgment: Teachers narrow AI scope using their expertise
• Privacy balance: Providing context while protecting student data

“We’re in this race for teachers to get this knowledge. This will become the most disruptive technology in our time. There is a real demand from educators to learn so that they are in the driver’s seat for AI.” — Randi Weingarten, AFT President

The Honest Take

This is the right approach. Teachers who understand AI’s capabilities can use it to augment their expertise rather than replace it. Agentic tools that stress-test lessons for gaps are genuinely useful—they apply a teacher’s judgment to find what might confuse students. The question is whether schools will invest in this level of training widely, or if only some teachers will get access.

2

Agentic AI Supports Real-Time Problem-Solving

March 20, 2026 | EdWeek

Teachers at the training are developing AI agents that help brainstorm alternative approaches when lessons or interventions aren’t working. For special education teachers like Yasheema Cook, this means creating agents that help monitor and adjust individualized education programs (IEPs).

• Intervention support: Paraeducators can brainstorm approaches on the fly
• IEP monitoring: Agents help track student progress and adjust plans
• Context injection: Teachers use expertise to guide AI toward relevant responses
• Hallucination reduction: Narrowing scope reduces AI errors

“A lot of teachers are doing this work at home, just wracking their brains trying to figure out what’s going to work for the next day. When something doesn’t work, it’s like, OK, now we’ve got to brainstorm on the fly, what else to do?” — Lois Torres, Preschool Paraeducator, NYC Public Schools

The Honest Take

This solves a real problem: teachers spending evenings trying to figure out what went wrong and what to try next. Agentic AI that can quickly suggest alternatives based on student context is genuinely valuable. The key is that teachers remain in control—they’re not outsourcing judgment, they’re accelerating their own problem-solving process.

3

Privacy Concerns Shape AI Implementation

March 20, 2026 | EdWeek

Teachers are carefully navigating the balance between providing enough context for AI to help and protecting student privacy. New York City public schools have not yet released formal AI use guidance planned for March, leaving teachers to make individual decisions about data sharing.

• Context vs. privacy: Balancing detailed student info with data protection
• Unclear guidance: NYC schools’ AI policy not yet finalized
• Tool variation: Different AI tools handle data differently
• Teacher judgment: Educators making individual decisions about AI use

The Honest Take

The privacy tension is real and unresolved. Teachers need context about student disabilities and challenges to get meaningful AI support—but that data is sensitive. Without clear district guidance, teachers are making these calls themselves, which creates inconsistency and risk. School districts need to provide frameworks now, not after problems emerge.

4

OpenAI Releases New Tools for Measuring AI Learning Outcomes

March 4, 2026 | OpenAI

OpenAI has released new tools for understanding AI and learning outcomes, advancing how AI’s impact is measured across learning environments. The initiative supports education as one of AI’s most promising frontiers with tools like ChatGPT.

• Learning measurement: Tools to assess AI’s educational impact
• Educational frontier: OpenAI positions education as key AI use case
• Evidence-based: Moving beyond anecdotal claims to measured outcomes
• Research support: Tools for researchers studying AI in education

What This Means for Educators

Agentic AI is becoming accessible to teachers. The National Academy for AI Instruction is training educators to build sophisticated tools—not just use pre-built ones. This matters because teachers can customize AI to their specific classroom needs.

The training gap remains significant. While 60% of teachers say they use AI, most still use it for basic tasks. Moving to sophisticated applications requires professional development that most districts haven’t funded.

Privacy frameworks are lagging behind practice. Teachers are making individual decisions about student data without clear guidance. Districts need to provide policies before problems emerge.

Share this article

𝕏 in ✉