The Center for Reinventing Public Education (CRPE) has released a new report on the pace and approaches to AI adoption by public education agencies across the country, and Virginia is among the states highlighted in the report.
Titled “Leading through Uncertainty: State Approaches to AI in Public Education,” the report is based on three data sets compiled between November 2025 and April 2026
- Survey responses from 29 State Education Agencies (in Virginia the SEA is the Virginia Department of Education) and 9 partner organizations
- Interviews with 14 SEA staff and leaders from 9 partner organizations
- Desk research conducted between November 2025 and January 2026 that led to a database and a brief of publicly available information on 20 Early Adopter states.
Together, the “Leading through Uncertainty” report includes information from 39 U.S. states and territories and 18 partner organizations.
Early on CRPE’s Radar
Virginia was recognized by CRPE as one of 20 “early adopter states” and included in its database and brief released in March of this year.
The database is nonevaluative, meaning CRPE was not attempting to rate the efficacy of how well these states were performing, but rather collecting key factual information around legislation and partnerships shaping AI policy.
The “Leading through Uncertainty” report, however, recognizes Virginia as one of three states actively investing in AI literacy and professional learning.
Along with North Carolina and Massachusetts, Virginia is recognized by CRPE not only for investing in literacy and professional learning, but for “jumping in early and working to scale.”
Singled out is Virginia’s online portal, Virtual Virginia, which is “developing AI-focused professional learning modules and micro-credentials for educators.”
Of note, according to CRPE, is that the state is adopting a “train-the-trainer model.” This allows the state to set a common baseline for teaching professionals then support broad access “so school divisions can adapt and deliver content locally.”
The strength of this approach is that it avoids the pitfalls of targeted guidance, which with a swiftly changing tool like AI can be rendered obsolete quickly.
By training the trainer, the state recognizes that AI implementation requires substantial human judgment, and that judgment rightly belongs in the hands of those most directly engaged in working it.
“Educators and leaders need opportunities to understand what AI tools can and cannot do,” the report writes, “how to evaluate outputs, how to protect student data, how to address bias and accessibility, and how to connect AI use to instructional goals.”
That was the point that Sharon Shewbridge made in an earlier Virginia Free Press article about AI in the classroom. “The missing piece” in implementing AI across a school district, the chief learning and innovation officer for Stafford County Public Schools told VFP, is “How are we teaching students ethical and effective use of artificial intelligence?”
Other Approaches
The report highlighted other innovative approaches to bringing AI into classrooms across the country.
One approach involves what CRPE calls “structured AI tool pilots.”
Currently being rolled out in Indiana and Louisiana, this approach has the state rolling out “AI tools in intentional ways.”
According to CRPE, “These pilots can support school systems in implementing AI thoughtfully, provide opportunities to learn from real-world use, and help states begin to understand whether specific tools are producing desired results.”
Another model that is more top-heavy is being adopted in Ohio, Tennessee, and Texas. Here, the state mandates policies rather than allowing local education agencies to experiment with models themselves.
“These approaches,” CRPE writes, “may help reduce confusion for districts and families by creating a baseline for acceptable use, privacy, and transparency.”
Challenges Ahead
Encouraging trends in AI adoption and use have a counterweight. Chief among these is an old problem in education – adding more to the plate of already overworked professionals.
“Few SEAs have a dedicated AI role or team,” CRPE reports. “AI work is often folded into existing responsibilities, even as staff are already managing other priorities.”
In Virginia, legislation requires an instructional technology and resource teacher (ITRT) for every thousand students. These individuals’ roles range from working with teachers and the district on adopted strategies for teaching AI use to working with and rolling out individual tools.
But even these roles don’t eliminate the concerns over staffing. Adam Garry, a consultant who works closely with many school districts across Virginia to understand and roll out effective AI strategies, told VFP earlier this year that a significant number of ITRT personnel are too focused on tools, which doesn’t help bring teachers along. That’s because tools training, he believes, “isn’t helpful.”
Another challenge is also a longstanding tension in education – respecting the difficult balance between state mandates and requirements, and the desire of local districts to act independently.
“Many SEA leaders said they lack the authority to mandate actions and the capacity to enforce them,” CERP reported. “They instead exert influence through guidance, frameworks, toolkits, and other ‘opt-in’ support. While helpful, these approaches could widen gaps among LEAs who opt in and those who do not.”
To date, Virginia looks to be one of the leaders in the country at finding a constructive path forward with AI and classroom deployment.
However, AI is moving faster than even districts eager to use it can keep pace. The answer lies not so much in finding answers, but supporting districts and collaborating.
“States do not need to have all the answers now,” CERP writes, “but they do need to create the conditions for better answers to emerge. … With stronger coordination and investment, states have an opportunity to help districts navigate AI responsibly and equitably. Without it, AI risks becoming another force that widens gaps rather than helping to close them.”

