AI's Classroom Gold Rush: Buying First, Asking Later

The education technology market has entered a strange new phase: school districts are spending heavily on artificial intelligence products, yet few can say with confidence which ones actually help children learn. The spending is real—budget lines for AI tutoring platforms, automated grading systems, and adaptive learning software have grown dramatically—but the evaluation process behind those purchases has not kept pace. Too often, the decision to buy is driven by a vendor's polished demo, a neighboring district's enthusiasm, or a superintendent's fear of being left behind, rather than by evidence of classroom impact.

A Procurement Pipeline Built on Hype

The structural reasons for this mismatch are not hard to find. Pilot programs are typically short, running a few weeks or a single semester, which is rarely enough time to measure genuine learning gains. Data from these pilots is often siloed within a single school or grade, never aggregated into a usable district-wide picture. Vendors, meanwhile, rarely publish independent efficacy studies, and the technology itself shifts so quickly that a product reviewed in the spring may be unrecognizable by fall. Most districts simply lack the staff—a dedicated edtech evaluator, a data analyst, a research partner—to do the rigorous vetting that a seven-figure contract deserves.

In New Mexico, the stakes are amplified by geography and scale. Rural districts, which make up a large share of the state's schools, contend with unreliable broadband, high teacher turnover, and administrative teams so small that a single technology director may oversee dozens of buildings. A tool that performs well in a well-connected suburban classroom can fail outright in a rural setting with spotty connectivity and limited technical support. When budgets are thin and every dollar is contested, a bad AI bet is not just wasteful—it is a direct subtraction from other priorities like staffing, facilities, and instructional materials.

The path forward is not to reject AI outright, nor to embrace it uncritically. What districts need are shared evaluation frameworks, regional purchasing consortia that pool expertise and bargaining power, and pilot designs that measure learning outcomes rather than engagement metrics like click rates or time-on-task. Until procurement catches up with the technology, the billions flowing into classrooms will continue to fund experiments—with students as the unwitting test subjects.