Key takeaways
-
Begin with a clearly defined educational or operational need. An AI tool should solve a real problem, not just add another piece of technology to a district.
-
Pilot the tool in your own district with real students, real work, and a small group of users.
-
Decide what success should look like before the pilot begins so you can judge whether the tool is worth expanding.
School leaders are facing growing pressure to make decisions about artificial intelligence. New products appear almost daily, staff members are experimenting with free tools, and vendors are making ambitious claims about saving time, personalizing learning, and improving results. Boards of education and families also want to know how schools are preparing students for a world in which AI will be common.
Before making a purchase, a district should understand how a tool will be used, what problems it is expected to solve, and what new challenges it may create. The goal is to choose technology that improves teaching, learning, or school operations and is worth the time, cost, and effort required to implement it well.
For me, thinking about what to consider before purchasing AI for schools starts with a simple question: What problem are we trying to solve? Once that is clear, districts are in a much better position to determine whether an AI tool actually addresses that need and will make a meaningful difference for students and teachers.
What Should Schools Consider Before Purchasing AI Tools?
When determining what to consider before purchasing AI for schools, leaders should examine the tool from several perspectives. Instructional leaders need to consider educational value, technology staff must review privacy and integration, and teachers and students can speak to usability. Business officials need to understand the full cost and contract terms.
That review should also account for the fact that AI is different from many traditional digital tools. Its outputs can change based on the user, the prompt, and updates to the underlying model. It can generate inaccurate or biased content with complete confidence. A feature that performs well during a vendor demonstration may respond differently when used by hundreds of students across a range of ages, needs, and reading levels.
Schools should ensure the tool is a good fit for the students and staff who will use it, and that people, not the technology, remain responsible for important decisions. If the tool gives feedback, evaluates student work, or helps shape decisions about a student, schools should be especially careful.
Does the Tool Solve a Real Problem and Align with Your Goals?
Before scheduling demonstrations, write a short problem statement. Teachers may need a faster way to create accessible first drafts of materials, or students may need timely practice and feedback in a specific skill.
The more specific the problem, the easier it becomes to evaluate the product. Ask the vendor to demonstrate the tool using realistic district tasks, content, and users. A polished generic demonstration may show what the product can do under ideal conditions. It does not show whether it will work within the district’s curriculum, technology systems, policies, staffing, or daily schedule.
Leaders should identify what should improve if the purchase is successful. Will teachers save measurable planning time? Will students receive better feedback? The measure does not need to be complicated, but it should be established before the purchase.
AI should also fit within the district’s broader approach to teaching and learning. A strongK-12 online learning platform can provide high-quality content and flexible ways for teachers to engage students. AI features should strengthen that work while preserving the role of the teacher, the integrity of the curriculum, and the thinking students need to do for themselves.
How Should Schools Evaluate AI Data Privacy and Security?
A privacy review cannot be reduced to merely locating the words “FERPA compliant” on a vendor website. Districts need to know what the tool collects, why it collects it, where it stores the information, how long it retains it, who can access it, and whether it uses it to train an AI model. They should also determine what happens to district information at the end of the contract.
To do this effectively, it’s helpful to start with rules established at the federal level. The U.S. Department of Education’sStudent Privacy Policy Office provides FERPA guidance for schools, and the Federal Trade Commission’sCOPPA guidance covers the online collection of personal information from children under 13. Districts also need to follow state laws, board policies, records-retention rules, and local purchasing requirements. For example, in my state, New York Education Law § 2-d includes additional requirements for protecting student data and personally identifiable information.
The contract matters, too. District leaders should ensure it clearly explains who owns the data, how it can be used, how it is protected, what happens in the event of a breach, how information is deleted, and whether the vendor can use the data for advertising or other commercial purposes. It is also important to ask how the vendor tests its security and fixes problems when they are found.
Privacy and security should be part of the decision from the beginning. If a district cannot clearly explain what data is being collected, how it is being used, and how it is being protected, it is not ready to require the use of the tool.
Why Do Teacher Training and Total Cost of Ownership Matter?
An AI tool can be easy to access yet difficult to use well. Teachers need time to understand where it adds value, where it can make mistakes, what information may be entered, and when its output requires closer review. They also need examples connected to their actual responsibilities, not training built entirely around the “perfect” prompts to use.
Professional developmentshould begin before the tool is rolled out and continue once staff start using it. Early training helps people understand how the tool should be used, while follow-up support gives teachers a chance to ask questions and work through problems that come up.
When it comes to cost, district leaders also need to look beyond the annual license fee. Other costs may include setup, rostering, data integration, staff training, substitute coverage, technical support, and future price increases. There is also a time cost to consider. Time spent implementing one product is time that cannot be spent on other district priorities.
Explore Educational Technology Resources
See how Discovery Education can support your school or district.
5 Questions to Ask Before You Buy an AI Tool for Your School
1. What Specific Problem Will This Solve?
Start by clearly identifying the problem the district is trying to solve and who the tool is intended to help. Be specific about what you want to improve and how the tool could help. If the main reason for purchasing it is simply that the features look impressive or other districts are using it, that is probably not enough reason to move forward.
2. Will the Tool Work in a Real School Setting?
Ask your vendors for more than testimonials. Take a look at who tested it, who participated, and what the results showed. Then try it in your own district with a small group of students, use real work, and set clear expectations for what success should look like.
3. What Could Go Wrong, and How Will People Remain in Control?
Test for inaccurate, biased, inappropriate, or inconsistent outputs. Determine how users report problems, how quickly the vendor responds, and which decisions always require educator review. TheNIST AI Risk Management Framework offers a useful structure for identifying and managing AI risk.
4. How Will the Tool Protect Student and Staff Data?
Do not rely only on a vendor’s word alone. Review the privacy policy, security information, and contract language. Make sure you understand what data the tool collects, how long it keeps that information, whether it is used to train the system, who else can access it, and how the data is deleted when it is no longer needed.
5. What Will It Take to Implement the Tool Successfully?
Look beyond the cost and consider what it will take to use the tool effectively. Decide who will introduce it, train staff, provide support, monitor use, and review the results. Set those expectations up front so the tool doesn’t become another subscription purchased but rarely used.
Making the Right AI Purchasing Decision
Schools do not need a complicated process for every AI tool, but they do need to evaluate each one consistently. A simple checklist can help define the need, review the evidence, check privacy and security, test accessibility and accuracy, and decide how to measure success.
Most importantly, test the tool on a real classroom or operational need before making a larger commitment. Listen to the people who will use it and to those responsible for privacy, security, and implementation. If the tool isn’t ready or doesn’t meet the district’s needs, don’t be afraid to walk away, even if the demonstration was impressive.
An AI purchase should make work easier, improve learning, and earn the trust of students, staff, and families.
