Schools do not need a static AI policy. They need a living AI framework that can be revised as tools, risks, and uses change. Utah’s approach is built around that reality: set conditions, define responsible use, protect student data, and keep professional development moving.
Build policy that can change
An AI framework has to be revisited often. New terms, new workflows, and new forms of AI can make old guidance obsolete quickly. A useful framework names both good uses and negative uses of AI, then stays open to revision as the field changes.
This matters because policy moves slowly, while AI moves fast. If schools want guidance that still makes sense a year from now, they need someone accountable for watching what changes and updating the document.
Center humans, not just tools
The strongest AI guidance keeps teaching and learning human-centered. AI should enhance the work of educators, not replace their judgment. That distinction matters.
Thinking of AI as a fundamental technology helps clarify the work. It is already present in many tools people use every day. The real question is not whether schools will use AI, but what they will plug into it:
- instructional strategy
- accessibility
- universal design for learning
- creative classroom practice
When those elements are strong, AI becomes a support for inclusion and empowerment.
Professional development cannot be one-and-done
Educators need ongoing AI literacy and AI fluency, not a single training session. Utah’s model uses repeated summits, teacher-facing training, and follow-up work that leads to lesson plans grounded in real classroom practice.
That kind of professional learning works because it is both structured and responsive. It lets educators learn the basics, try ideas in context, and return with questions. It also helps build a community of practice so people are not trying to keep up alone.
Data privacy has to be built in
AI use in schools must be paired with strong data privacy safeguards. Student data privacy agreements are a practical step toward making sure products meet expectations before they enter school systems.
That is especially important because AI systems can absorb information in ways families and educators may not fully see. Leaders should know what tools are on their servers, what filters are in place, and what teachers and students are being asked to log into.
If a company refuses to sign a data privacy agreement, that is a warning sign. Schools should not treat privacy as optional.
Get ready for agentic AI carefully
Agentic AI refers to multiple AI systems working together toward a task. It is a meaningful next step, but most districts are not ready to jump there yet. They still need stronger fluency with the basics before moving into managing agents.
That transition will also require clearer thinking about roles. Educators will need to move from using tools to managing ecosystems, and that is a major shift in skill and responsibility.
Families should be part of that preparation too. If parents are left out, they will get their information from sensational headlines instead of grounded school communication. Schools that want trust around AI need transparency and family engagement, not just internal training.
The bottom line for school leaders
A practical AI framework should do four things well:
- stay current through regular revision
- support responsible use with clear guardrails
- invest in ongoing professional development
- protect student data through formal privacy agreements
Schools do not need to chase every new AI feature. They need a disciplined framework that helps them use AI carefully, creatively, and in service of teaching and learning.
Frequently Asked Questions
What makes an AI framework effective for schools?
It must be a living document that gets revised often, names good and negative uses of AI, and stays current as the field changes.
Why is ongoing professional development important for AI?
Because AI keeps changing, educators need repeated training to move from basic literacy to real fluency and classroom use.
How should schools approach data privacy with AI?
They should require data privacy agreements, check what tools are in use, and block products that will not meet privacy expectations.
Are districts ready for agentic AI?
Not fully. They still need stronger AI fluency first, and families should be included before schools move deeper into agentic workflows.
For more context, listen to the original episode of #PrincipalPLN: The Learning Experience: What Utah is Building with AI & Why it Matters.