Rather than approaching AI as a traditional technology purchase, some districts are reframing the challenge as "adaptive governance" — a concept detailed in a new industry white paper, Adaptive Governance for AI in K-12. This model focuses on building the organizational competency to manage a technology that evolves monthly, not annually.
“This is not just something that my systems and technology team is going to be thinking about,” Dr. Aleesia Johnson, superintendent of Indianapolis Public Schools, said in a webinar Monday. “This is something that’s going to impact us across the organization.”
Education leaders emphasized that AI represents a departure from previous technology waves like learning management systems or personalized learning dashboards. According to Johnson, two factors create this unique urgency: the unprecedented rate of change, and AI’s ability to create work products rather than just supporting existing workflows.
Because AI impacts students, staff and families simultaneously, district leaders said they could no longer simply delegate AI strategy to IT departments. They said effective implementation now requires superintendents to be personally proximate to the decision-making process to ensure AI aligns with broader organizational goals.
Moreover, speakers emphasized that AI implementation is fundamentally a leadership issue, unlike traditional technology rollouts. Dr. Julia Rafal-Baer, CEO of education strategy and policy firm ILO Group, who also moderated the discussion, framed the shift as one from tool adoption to system design, where leaders must build the capacity to “make decisions, learn quickly and adapt over time.”
District leaders participating in the webinar described moving away from two common approaches: ignoring AI altogether or allowing widespread, uncoordinated experimentation.
In North Carolina’s Charlotte-Mecklenburg Schools, Superintendent Crystal Hill said she took a more deliberate “third path,” centered on stakeholder engagement, AI governance and phased implementation. Before introducing any AI tools, the district began by collecting feedback from roughly 10,000 students, families and staff on everything from student interest in how AI systems work to family concerns about safety to teacher demands for professional development. That input informed the development of policies, governance structures and cybersecurity safeguards before any tools were deployed, Hill said. The district then launched targeted pilots in 30 schools, requiring participants to define specific problems, document implementation and measure impact.
“We wanted to focus on pilots … instead of turning on a lot of different tools or letting people do their own thing,” Hill said.
Another shift happening in some districts, described by participants as a key component of the new governance model, is the move away from product-specific training toward building general internal capacity for AI use.
For example, Johnson said that rather than training staff on specific platforms, Indianapolis Public Schools is focusing on developing foundational AI skills like prompt engineering.
“If you know how to be a strong prompt engineer, that is sort of agnostic to a platform that you're choosing,” she said, adding that this approach transforms the district from a passive recipient of vendor pitches into an informed buyer capable of solving local problems. It’s also a strategy the ILO Group report identified as essential for maintaining control over AI adoption in a rapidly changing market.
Across four speakers at Monday's webinar, one principle emerged consistently: AI efforts should begin with clearly defined problems, not available products. In Charlotte-Mecklenburg, Hill said schools participating in pilot programs were required to identify areas of friction — such as time-consuming tasks or gaps in communication — before testing AI solutions. The district then evaluated results based on measurable outcomes, including time savings, quality improvements and overall impact.
Cindy Marten, Delaware’s secretary of education and former U.S. deputy secretary of education, also warned that without this disciplined focus, innovation risks becoming mere "noise."
“If it doesn’t make learning stronger, safer or fairer, it’s just noise,” Marten said, quoting Delaware Gov. Matt Meyer.
Marten said she views the state’s role not as a gatekeeper that slows down innovation, but as the builder of a "superhighway." By providing clear guardrails, safety guidance and procurement support, she said, states can allow districts to move quickly without risking student data privacy.
Marten added that Delaware is developing an "assurance lab" to evaluate tools and share best practices, with the goal of sparing the state's 19 school districts from having to do it themselves and reducing the risk of inconsistent adoption, and thereby inequitable outcomes for students.
“Start with the problem, not with the tool,” she said.
To prevent "decision gridlock" caused by competing concerns from legal, IT and instructional teams, district leaders recommended bringing all parties to a shared governance table early.
“Bicycles don’t go on freeways,” Marten said, suggesting that clear terms and conditions allow districts to reject tools that don't fit safety criteria without endless debate.
Ultimately, district leaders participating in the webinar said the goal of adaptive governance is not to get AI "perfect" the first time, but to build the organizational muscle to decide, learn and adjust as the technology continues to shift the educational landscape.
