Schools are buying AI fast, but their contracts still can’t show it helps

Utah’s statewide screen-time law took effect July 1, and Los Angeles Unified School District has approved a stricter local screen-time policy for the 2026, 27 year, according to THE Journal. Those moves are arriving during the same purchasing season when districts are adding AI tools in classrooms, central offices, and student-support operations, yet many still do not spell out in contract terms what counts as success.

That disconnect is the core operating challenge in 2026. The question is not whether AI belongs in schools. It is whether districts can procure it in ways that can withstand audits, board review, and renewal decisions.

AI budgets are real, but buying standards are still unclear

Education Week reported Aug. 20 that district officials are swamped by the volume of AI learning tools and by how quickly vendors release updates. The piece, by Robbie Sequeira of Stateline.org, depicts districts weighing which products are safe and worth spending thousands or millions of dollars on, while state and local guidance for reviewing purchases varies widely.

Part of the intensity comes from the market’s scale. Education Week, citing Grand View Research, reported that the wider U.S. education technology market generated nearly $48 billion in 2024 and could exceed $90 billion by 2030.

Education Week also described a common public-sector imbalance: vendors usually understand their products far better than districts do, so districts have to create their own ways to evaluate what they are buying. And even when a district knows its goal, it can be difficult to translate that goal into requirements when “AI” is sold as an add-on feature for almost any product.

Education Week offered a district example to illustrate the purchasing challenge. It described Sumner County Schools in Tennessee testing an AI-powered literacy tool. Beyond that, this article does not add further details about the district’s reading trends or leaders’ specific in-class evaluation focus, because those claims were not supported by the sources provided.

Screen-time policies are changing what districts must document

In an Aug. 25 viewpoint, THE Journal argues that simply tallying total student screen time misses the more important question, and that leaders should judge the value of each individual ed tech program. In practice, whether districts favor the policy or not, limits set by a state or a major district push programs to compete for fewer instructional minutes.

THE Journal described Utah as the first statewide policy setting age-based expectations for K-12 technology use, effective July 1, and it noted LAUSD’s tougher policy for the coming school year. For CIOs and instructional leaders, the impact is straightforward: rules framed around “time” quickly turn into arguments over which tools earn that time, and those arguments require evidence and reporting that many older edtech contracts never demanded.

THE Journal also emphasized that accountability will be the hardest part next, expecting school leaders to hold providers to higher standards and to make sure products are shown to help students learn. That becomes a procurement and data issue as much as a teaching issue. When vendors cannot supply usable proof, districts will need to generate their own, which requires usage tracking, baseline measurement, and checks on implementation fidelity.

AI is turning into a workflow, and integration costs rise with it

In the classroom, AI use is moving from one destination site to many task-specific tools, according to eSchool News (Aug. 17). The outlet describes teachers turning to different AI platforms to brainstorm, tailor materials, create formative assessments, translate messages, and review student work. It urges educators to build an “AI workflow” instead of collecting an “AI toolbox.”

For IT teams, that shift matters because workflows create integration work. More tools often bring more accounts, more data exchanges, more permission settings, and more chances for uneven oversight. Even districts with a standard LMS and SIS can end up with AI products operating outside those controls unless purchasing requirements include SSO, rostering, and administrator reporting from the beginning.

Tech & Learning’s essay about what a parent learned from a freshman’s everyday AI use makes a similar operational point: student practices can change faster than policy, and the system that matters is what students and teachers actually do, not what the acceptable-use policy says. That mismatch increases the need for districts to offer clear, approved ways to use AI in class.

What to include in next year’s AI RFPs and renewals

Education Week’s reporting points to the market reality: large sums are moving into AI for education, while districts still shoulder much of the evaluation work. THE Journal’s screen-time developments add pressure that will surface in board discussions and parent communications. eSchool News, meanwhile, describes classroom adoption that is spread across many tools rather than centered on one platform.

For district operators, that mix suggests AI purchasing is heading into a stage other enterprise software categories reached earlier, where value claims must be specific, measurable, and defensible in an audit. Brookings Institution commentary on “making AI work for schools” (July 31, 2025) also underscored responsible integration. While it is not a 2026 news development, it can still inform planning for districts building governance now.

Procurement checks to run before you sign or renew

• Define the job before you buy the tool: state the academic or operational problem in one sentence, then require the vendor to tie features, rollout steps, and success measures to that problem, consistent with the needs-first framing described in Education Week’s district example.
• Require per-application usage reporting that administrators can export: adoption by school, grade, course, and teacher, plus patterns tied to the instructional model. Screen-time policies such as Utah’s and LAUSD’s, as cited by THE Journal, make this documentation a defensive requirement.
• Treat “workflow sprawl” as an integration requirement: require SSO, rostering, and clear data-sharing terms for every AI product that touches student information, since eSchool News describes teachers using multiple tools for different tasks and governance can slip when each purchase is handled as a one-off.
• Include professional learning in the implementation plan and budget: eSchool News stresses using AI thoughtfully to free time for human teaching work. Districts should budget training hours with milestones, not as an optional add-on that disappears when schedules tighten.