Justin Reich, director of the MIT Teaching Systems Lab and a leading researcher in education technology, has issued a pointed warning to schools rushing to adopt artificial intelligence: many are moving faster than the evidence allows. In a June 2026 analysis, Reich argues that schools should be wary of unproven AI initiatives and rethink how they evaluate new technologies before committing resources and student time.
The warning comes at a moment when AI adoption in K-12 and higher education is accelerating faster than any previous technology wave. From adaptive learning platforms to AI tutors and automated grading systems, institutions are under pressure to demonstrate AI readiness. Reich's message is that readiness should not be measured by procurement speed, but by the quality of evidence behind the tools.
The Evidence Gap
Reich points to a persistent gap between the capabilities vendors promise and the outcomes researchers can verify. Many AI education tools claim personalization, engagement, and learning gains, but few have undergone rigorous independent evaluation. The risk, Reich argues, is that schools spend scarce resources on tools that sound transformative but produce marginal or even negative effects on student learning.
For school leaders, the challenge is not whether to adopt AI, but how to separate genuine innovation from marketing hype. Reich advocates for structured pilot programs with clear evaluation criteria, not district-wide rollouts based on vendor demos or conference presentations.
What Schools Should Do Instead
Rather than chasing the newest AI tool, Reich suggests schools should invest in what research already supports: teacher professional development, structured feedback systems, and curriculum design that integrates technology as a support rather than a replacement for human instruction. The schools that will benefit most from AI, he argues, are those that already understand what they are trying to teach and how they will know if it works.
For ed-tech vendors, Reich's warning is a market signal: the window for selling unproven AI tools is narrowing. Schools are increasingly demanding evidence, and the vendors who can provide rigorous research backing will win the next procurement cycle.