The Tool Was Approved. Was the New AI Feature?

One of the assumptions built into school technology approval is that the product a district reviews will remain reasonably close to the product teachers eventually use. A platform goes through a privacy review, someone considers the instructional purpose, the district decides whether it belongs on the approved list, and teachers are told they can use it.

AI is making that process less stable because many of the tools already inside schools are changing after the approval happens.

A platform that was originally used for writing practice may add an AI assistant. An assessment tool may begin generating feedback automatically. A learning platform may introduce a chatbot, automated recommendations, or a feature that allows students to generate text. The product name stays the same, the district approval stays the same, and the tool may continue appearing on the same list teachers have relied on for years.

What changed is what the tool can now do.

That creates a problem because teachers generally do not separate the platform from every feature inside it. If the district says a tool is approved, most educators reasonably assume the tool they open on Tuesday is still covered by the decision the district made months earlier.

A teacher may sit down to prepare an assignment in a platform they have used for several years and notice that students can now ask the system to generate a response, revise their writing, or provide feedback. The teacher may not know whether the feature was part of the original review, whether it can be disabled, or whether the district has issued any guidance about how students should use it.

The teacher still has to make a decision.

They can allow the feature because the platform is approved. They can tell students not to use it because they are unsure. They can stop using that part of the platform until someone answers the question. They can contact a technology administrator and hope the answer comes before the lesson begins.

None of those choices necessarily reflects poor judgment. The problem is that the teacher has been given an approved tool without clear information about whether the approval still reflects the tool in front of them.

The same issue appears when a principal receives a question from a family.

A parent may notice that their child is using an AI feature inside a district platform and ask what the feature does, what information it uses, or whether students are supposed to have access to it. The principal may know the platform is approved because it has been used across the school for years.

That does not mean the principal knows when the new feature appeared.

The question may move from the principal to the district technology office, then to curriculum, privacy, procurement, or the vendor. Someone may need to review the original agreement, determine whether the feature is covered, and decide whether staff need new guidance.

What looked like a straightforward question becomes more complicated because the school is trying to determine whether the product it approved is still functioning under the same conditions.

District technology teams are dealing with this from another direction.

They may review hundreds of products used across schools, departments, grade levels, and programs. Some of those products update every few weeks. Others introduce AI gradually, beginning with an optional teacher feature and later adding student-facing functions, automated recommendations, or new ways of processing information.

It is not realistic to expect a district employee to notice every change the moment it appears.

At the same time, it is difficult to tell teachers that the approved list is the boundary for responsible use if the products on that list can change in ways the district has not yet examined.

Schools are still treating technology approval as a decision about a product, even though AI is increasingly making approval an ongoing question about what that product has become.

This becomes especially difficult because not every update matters in the same way.

A platform may change its navigation, add a reporting feature, or reorganize how teachers access materials without creating a meaningful new concern. Another update may introduce a student-facing AI assistant that changes how students complete work, what information the system receives, or what kind of support the platform provides.

Both are software updates.

They do not require the same response.

The challenge for schools is deciding which changes are routine and which ones alter the conditions under which the original approval was given.

A curriculum director may face this during an instructional review. Teachers are using a platform exactly as expected, but a new automated feedback feature begins suggesting revisions to student writing. The district originally approved the program because it supported practice and allowed teachers to monitor progress.

The new feature may still be useful. It may help students identify errors more quickly or give teachers another way to provide support.

But it may also change the instructional question. If students are now receiving generated feedback before the teacher sees the work, the district may need to decide what role that feedback should play, whether it aligns with instructional expectations, and whether students understand the difference between automated suggestions and teacher guidance.

The fact that the platform was already approved does not answer those questions.

The same distinction appears in student support systems.

A school may use a platform that organizes intervention data or helps staff track progress. If the vendor later adds an AI feature that begins recommending next steps, the system has moved from storing information to interpreting it.

That may be a useful addition.

It is also a different function.

The school now has to consider how those recommendations should be used, what information informs them, and whether educators understand the limits of the suggestion before it becomes part of a decision about a student.

The change may seem small from the vendor’s perspective because it exists inside the same product. For the school, it may create an entirely different set of questions.

Why It Matters

Teachers rely on approval systems because they cannot reasonably investigate every technology product themselves. A classroom teacher should not need to read vendor agreements, monitor release notes, interpret privacy policies, or determine whether a newly introduced AI feature changes the conditions under which a platform was originally approved.

That is not the work teachers were hired to do.

When the approval process does not keep pace with the product, that work begins moving toward the classroom anyway.

Teachers start deciding whether a feature is appropriate because they are the first people who see it. Principals start answering questions because families bring concerns to the school. Curriculum leaders begin interpreting new functions because they affect instruction. Technology teams are asked to review changes after the feature is already being used.

The burden is spread across the system, but the responsibility is not always clear.

That uncertainty can create inconsistent practice.

One teacher may use a new AI feature because the platform is on the approved list. Another teacher in the same district may turn the feature off because no one explicitly said it was allowed. A principal may tell staff to wait for guidance, while another school continues using the tool because there has been no formal restriction.

The district can have one approved platform and several different interpretations of what approval means.

This also affects trust.

When leaders tell educators to use approved tools, teachers need to know that the phrase carries enough meaning to guide their decisions. If they repeatedly discover features that no one has explained, they may begin to treat the approved list as incomplete information rather than a reliable boundary.

Families may experience the same uncertainty from the outside. A parent who asks whether an AI feature has been reviewed should not receive different answers depending on which person happens to respond.

The issue is not that schools should prevent products from changing. Technology will change.

The issue is whether schools have a way to notice when the change is significant enough that the original decision needs another look.

Schools can begin by separating product approval from feature approval more clearly.

District technology and curriculum leaders can review the current approval process together and identify the kinds of changes that should trigger another review. This does not need to include every software update. A change in navigation or reporting may not require much attention, while a new generative feature, student-facing chatbot, automated recommendation system, or function that processes student information may deserve a different level of review.

The important part is deciding that distinction before a problem occurs.

Technology and procurement leaders can also address this during vendor purchasing and renewal conversations. Along with asking what a product currently does, districts can ask how they will be notified when significant AI capabilities are added, whether those features can be disabled, and whether the district will have an opportunity to review them before they become available to students or staff.

That moves part of the responsibility back to the vendor instead of leaving the district to discover changes on its own.

School leaders can make the process clearer for teachers at the building level.

If an educator notices a new AI feature inside an approved platform, there should be one clear place to send the question. The teacher should not have to decide whether the feature is safe, whether the privacy agreement covers it, or whether the district intended for students to use it.

The teacher’s responsibility can be to notice and report. The district’s responsibility can be to review and respond.

That distinction matters because teachers already make hundreds of decisions about instruction, student needs, assessment, behavior, and support. Technology approval should reduce unnecessary uncertainty rather than add another category of decisions they are expected to make alone.

Curriculum and instructional leaders can also build this into conversations that are already happening.

During grade-level meetings, department meetings, coaching conversations, or curriculum reviews, leaders can ask whether the platforms teachers use most often have introduced any new AI functions. The goal is not to turn teachers into product monitors. It is to create a simple way for changes to become visible before they spread quietly across classrooms.

Principals can do something similar when staff raise questions about a feature.

Instead of asking the teacher to interpret whether the feature is covered by existing guidance, the leader can treat the question as information the system needs to examine. That sends an important message to educators: noticing uncertainty is not a failure to understand the technology. It is part of using it responsibly.

Districts can also become more specific in the language they use when communicating approval.

A tool may be approved for privacy and security without every instructional use being endorsed. A teacher-facing AI feature may be permitted while a student-facing function is still under review. A product may be approved for one grade span or purpose and not another.

Those distinctions may feel more complicated than placing a tool on a single list, but they can actually make the work clearer because teachers know what the approval is intended to tell them.

The goal is not to create a perfect system that catches every change immediately. That is probably not realistic.

The goal is to make sure the school has a reasonable way to identify when a tool has changed enough that the original assumptions should be revisited.

Teachers need to be able to rely on the systems around them. If a district asks educators to stay within approved technology, then approval has to remain meaningful after the purchasing decision is over.

Leaders need the same thing. They should be able to answer staff and family questions without first trying to reconstruct how a platform changed and whether anyone knew.

AI is making many familiar technologies less fixed than schools are accustomed to. The platform a district reviews in September may not be functionally identical to the platform teachers are using in February.

That does not mean schools should stop approving technology or become suspicious of every update.

It means approval has to become more responsive to change.

The question is no longer only whether a school chose the right tool.

It is whether the school still understands the tool it chose.

This is one of many hidden shifts AI is introducing into education. I write these newsletters to help educators and school leaders see them early and respond intentionally. If that matters to you, stay with the series.