Generative Artificial Intelligence (AI) has arrived in K-12 education quickly—and it’s not going away. As a Special Education Director who spends a lot of time in conferences, webinars, and meetings with families, I’m hearing the same questions across districts: “Is this safe?” “Is it cheating?” “Can it help students with disabilities?” “Do we ban it or build guidance?”
This post is designed to be an easy-to-read, practical guide based on the Oregon Department of Education’s resource on Generative AI in K-12 classrooms. I’ll focus on what matters most for district leaders, educators, and school partners (including online therapy providers like TinyEYE): clarity, equity, student privacy, and smart implementation.
What is Generative AI (in plain terms)?
Artificial intelligence is a broad field. Generative AI is a specific type that can create new content—most commonly text—based on a prompt. Tools like ChatGPT, Bard, and Microsoft Copilot use “large language models” trained on massive datasets (often pulled from the internet) to generate human-like responses.
Two key reminders for schools:
- Generative AI tools can sound confident even when they are wrong. They are not “thinking” like a person; they are predicting likely words based on patterns.
- Because they are trained on internet-based data, they can reflect the biases and inaccuracies found online.
I often use a metaphor shared in national education technology guidance: we want AI to function more like an electric bike than a robot vacuum. In other words, the educator stays fully in control, and the tool reduces burden and extends capacity—but doesn’t replace professional judgment.
Why talk about AI now?
Schools have used AI-adjacent tools for years (translation tools, adaptive learning programs, Grammarly-style writing supports). What’s different now is the speed, accessibility, and power of generative AI. Students and staff can use it instantly, at home or on a phone, often without realizing the implications for privacy, academic integrity, or equity.
That’s why districts need proactive planning. When policy lags behind practice, we end up reacting to incidents instead of building safe, consistent systems.
Equity implications: what districts must keep front and center
Generative AI sits inside a larger “digital learning ecosystem.” If we implement it thoughtfully, it can increase access. If we implement it carelessly, it can widen gaps. The ODE resource highlights several equity concerns that are especially relevant for K-12:
1) Bias
Bias can show up in many ways: privileging certain dialects or language structures, reinforcing racial or gender stereotypes, presenting a U.S.-centric viewpoint, or missing cultural context entirely.
Practical district moves:
- Train staff to recognize bias and teach students to question AI outputs.
- Position AI as a starting point, not an authority—students must verify information.
- Offer family-friendly resources so caregivers understand risks and limitations.
2) Inaccuracy (including “AI hallucinations”)
Generative AI can fabricate sources or present incorrect information in a believable way. That’s not a rare glitch—it’s a known limitation.
Practical district moves:
- Teach lateral reading and fact-checking as core digital literacy skills.
- Require students to corroborate AI-generated claims with credible sources.
- Provide staff with tools and routines for evaluating AI-supported work.
3) Plagiarism and academic integrity
Yes, students can use AI to write assignments. But bans alone rarely solve the underlying issue. Many districts are shifting toward clearer expectations and better assignment design.
Practical district moves:
- Rethink tasks to emphasize process: drafts, reflections, conferences, and in-class checkpoints.
- Increase formative assessment so teachers recognize each student’s writing “fingerprint” over time.
- Use AI detectors cautiously. They can produce false positives—especially for students who write formally or use language supports.
- Teach ethical use: when AI is allowed, how to cite it, and how to explain how it was used.
4) Copyright and licensing unknowns
Because generative AI is evolving rapidly, copyright boundaries and ownership of AI-generated content can be unclear.
Practical district moves:
- Review Creative Commons licensing basics with staff.
- Coordinate with district legal counsel and technology leadership when adopting tools.
- Clarify expectations for staff-created materials and student-created products.
5) Equity of access
Access is not just about devices—it’s also about training, language, disability supports, and future readiness. If some students learn to use AI responsibly while others are blocked from it entirely, we may unintentionally create a new “AI literacy gap.”
Practical district moves:
- Include students with disabilities, multilingual learners, and underserved communities in planning conversations.
- Watch for future paywalls and subscription models that could limit access.
- Invest in staff learning so access is meaningful, not merely available.
Student data privacy: the non-negotiables
When schools consider AI tools, privacy must be addressed upfront. Federal and state requirements may include:
- FERPA (Family Educational Rights & Privacy Act)
- CIPA (Children’s Internet Protection Act)
- COPPA (Children’s Online Privacy Protection Act), especially for children under 13
- State-specific protections (for Oregon, OSIPA is highlighted in the ODE resource)
One of the biggest practical risks is that prompts entered into generative AI systems may be stored and used to improve the model. That means staff and students should not input personally identifiable student information, disability information, IEP details, or sensitive family data into public AI tools unless the district has vetted the platform and established clear agreements and safeguards.
District leaders should:
- Review vendor privacy policies and terms of service regularly (not just once).
- Confirm age requirements and parental consent needs under COPPA.
- Align AI guidance with existing technology policies and cyber safety plans.
- Train staff on what data must never be entered into an AI prompt.
For partners like TinyEYE supporting schools through online therapy services, this is also a critical alignment point: therapy documentation, student goals, and service notes are highly sensitive. Any AI-enabled workflow must be vetted through the same privacy and compliance lens districts apply to all student data systems.
The potential: how generative AI can support learning (when used well)
When implemented with strong guardrails, generative AI can support educators and students in meaningful ways:
- Learning design: Drafting lesson ideas, creating question sets, or generating UDL-aligned options—always with the educator as the final decision-maker.
- Instructional support and differentiation: Rewriting text at different reading levels, generating vocabulary supports, or offering translated drafts for multilingual learners.
- Virtual assistant tasks: Drafting communications, summarizing information, or organizing ideas—freeing time for relationship-building and instruction.
- Student support and guidance: Helping students practice asking better research questions, outlining writing, or exploring career pathways (with adult oversight).
- Future readiness: Building AI literacy and digital citizenship so students can participate ethically and effectively in college and careers.
Special education and related services: a practical lens
In special education, we constantly balance access, independence, and dignity. Used carefully, AI can support:
- Text-to-speech and speech-to-text workflows that reduce barriers to reading and writing
- Language practice opportunities for multilingual learners
- Scaffolding and simplified explanations that help students access grade-level concepts
At the same time, we must be clear: generative AI is not a clinician, not a teacher, and not an IEP team member. It cannot replace professional expertise or the relational work that drives student progress. And because of privacy laws and ethical obligations, we must be especially cautious about what information is entered into AI tools.
This is where staffing realities matter too. Many districts are facing therapist shortages (speech-language pathologists, occupational therapists, school psychologists). Online therapy partners like TinyEYE can help districts maintain service delivery, but AI tools should be viewed as supportive infrastructure—not as a replacement for qualified providers or legally compliant IEP services.
What districts should consider when developing AI policy
District policy doesn’t have to be perfect on day one, but it must be clear, equity-centered, and actionable. Consider these starting points:
- Define allowed vs. not allowed use by grade band and context (instruction, homework, assessment).
- Clarify academic integrity expectations including citation guidance and what “ethical use” looks like.
- Build staff capacity through professional learning focused on bias, accuracy, and instructional design.
- Communicate transparently with families so expectations are consistent across school and home.
- Align with the district’s digital learning ecosystem (relationships, mental health, high-quality materials, readiness, and funding).
- Keep humans in the loop—educators remain responsible for decisions, evaluation, and student support.
Ultimately, the goal is not to chase every new tool. The goal is to protect students, support educators, and prepare learners for a world where AI will be present in higher education, the workforce, and daily life.
For more information, please follow this link.
