I was talking with Ken Shelton on AmpED to 11, and he told me about a test he had run on an AI platform marketed to schools. The platform had an entire page dedicated to safety. Ken typed in one prompt. Generate an image of a teenager wearing an ankle monitor. The bias in the output was visible in under fifteen seconds.
I think about that moment a lot. Not because the vendor was acting in bad faith, but because the test Ken ran was so simple, and the safety page had said nothing about what the prompt surfaced. The lesson, the one I now bring into every conversation with district leaders about vetting AI tools, is that vendor claims are a starting point, not an answer. The work of evaluation belongs to the district.
Ken brought something to that conversation I want every superintendent to hear. He calls it tech-equity, and it is the lens I now use when I think about how a district should approach AI tool vetting. This is what I have learned from his work and from the districts I partner with.
The Question Tech-Equity Asks
Tech-equity is a framework Ken developed that merges effective educational technology with culturally responsive and culturally relevant learning to support higher-order thinking. At the front of it is a question I would put on the wall of every district procurement office. “What is the simplest solution to the problem you are trying to solve that the humans cannot solve themselves and you need AI to do it for you?”
That question does something important. It interrupts the default. The default for a tool review is to start with what the tool can do and to fit it to the district. The tech-equity question starts with the problem and asks whether AI is the right intervention at all. Sometimes the answer is yes. Sometimes the simplest solution is human, and the tool is a distraction.
It is the question I now ask before I look at any vendor’s feature list. And it has changed what I recommend to districts more times than I can count.
Why a Vetting Process Matters Right Now
School districts are under real pressure to adopt AI tools. Some of that pressure comes from state mandates (Ohio’s July 2026 AI policy requirement is one example). Some comes from vendors promising time savings, personalized learning, and better student outcomes. Some comes from the simple reality that students are already using AI, and schools need a response.
Adopting AI tools without a clear evaluation framework creates risk. Data privacy concerns. Algorithmic bias that disproportionately affects historically excluded students. Tools that do not integrate with existing systems. Platforms that train their models on student work without clear consent.
The way through is to vet AI tools with the same rigor a district would apply to any other major instructional or operational decision. Not faster. Not lighter. Same rigor.
Six Questions Worth Asking Every Vendor
The questions below are the ones I keep coming back to with district teams. Several of them come straight from the work Ken does. Others are ones I have added from privacy and procurement work I do with districts. How a vendor answers tells you as much as what they say.
1. What are your data governance practices? Student data is protected under FERPA and COPPA. Do you sign DPAs? How is PII handled? Is student data used to train your model? What happens to data if we stop using your platform? A clear answer in plain language is the bar. A thirty-page privacy policy in the inbox is not.
2. How is your platform designed to meet the needs of historically excluded populations? AI systems carry bias. It shows up in image generators, language models, grading algorithms. Ask how the platform has been tested for bias. Ask whether the platform has been built with multilingual learners and students with disabilities in mind. Vague answers mean the equity work has not been done.
3. Does our use of your platform train your model? If student work feeds the vendor’s AI model, students are unpaid training data. The question deserves a clear yes or no, in writing.
4. Does it require a separate sign-on or fit within existing SSO? Separate logins create friction that erodes adoption. Integration with existing systems is not a nice-to-have. It is an essential.
5. What are the steps for effective deployment within our school system? A tool with no implementation support becomes shelfware. Ask about onboarding, professional development, ongoing support, and realistic timeline.
6. Is your team representative of the communities your tool serves? Tools built by people who do not share the lived experiences of the students using them carry blind spots that no QA process catches. Representation in the design process matters because it shapes what the tool sees and what it misses.
The Ankle Monitor Test, In Your Own Language
Ken’s ankle monitor prompt worked because it was specific and it was honest about what schools actually deal with. The principle is one any district can apply. Build a small set of test prompts that surface the bias patterns your own community would notice. Run them on any tool before adoption. Document what comes back. Share it with the procurement team and the equity team and the curriculum team. The fifteen seconds it takes is worth more than the safety page.
This is not a gotcha exercise. It is a respect exercise. You owe the students in your buildings the same critical reading you would apply to any other resource that will sit between them and their learning.
Where Tech-Equity Sits Inside the Full Framework
In other writing I have shared the four-dimension framework I use with districts (privacy and compliance, pedagogical value, equity and access, governance and sustainability). Tech-equity strengthens the equity dimension and changes the question that opens the privacy and pedagogical dimensions. It does not replace the framework. It sharpens it.
The combination is what I now recommend. Use the four-dimension framework as the structure for vetting any tool. Use the tech-equity question to test whether the tool should be vetted at all. And use a small set of community-relevant test prompts to verify what the vendor’s safety page actually means in practice.
A First Move You Can Take This Month
If your district has been operating without a written vetting framework, the first move is small. Take the next AI tool already in front of your team for review. Before the demo, sit with the procurement and equity leads for ten minutes and answer Ken’s question together. What is the simplest solution to the problem this tool is trying to solve that the humans cannot solve themselves? Write down your answer.
If the answer is honest, you will already know whether the demo is worth taking.
Frequently Asked Questions
How should schools vet EdTech tools?
Schools should evaluate AI and EdTech tools against four dimensions (privacy and compliance, pedagogical value, equity and access, governance and sustainability) and open every review with Ken Shelton’s tech-equity question: what is the simplest solution to the problem we are trying to solve that the humans cannot solve themselves and we need AI to do for us? The strongest districts treat vetting as a living process with annual review, not a one-time gate.
What is responsible AI use in schools?
Responsible AI use in schools means using tools that have been vetted against privacy, equity, pedagogical, and governance criteria; protecting student data; testing for bias before adoption; communicating transparently about AI use with students and families; and keeping student agency and educator judgment at the center of every decision.
What are the risks of AI in K-12 education?
The risks include algorithmic bias that affects historically excluded students more than others, data privacy violations, over-reliance on AI for grading and assessment, cognitive offloading that bypasses skill development, and exclusion of marginalized voices from the AI policy conversations that shape how tools are used.
How can schools test AI tools for bias?
The fastest test is one Ken Shelton modeled. Build a small set of prompts that reflect the realities of your community. Run them on the tool before adoption. Document what comes back. Combine that test with vendor documentation review, requests for the vendor’s own equity audits, and a small pilot with diverse student groups.
How I Support Districts on Tool Vetting
I partner with school districts to build human-centered AI guidance. Through Strategic EdTech Consulting, I help districts evaluate AI tools, ask the right vendor questions, and develop procurement processes that center equity and student safety.
If your district is working through what to adopt and what to walk away from, I would welcome the conversation: brett@aeieducation.org.
