School districts are rushing to adopt generative artificial intelligence tools to plan lessons and grade assignments. Educational experts warn that without a clear problem to solve, these high-tech investments usually fail. Research shows that unstructured AI tools can hurt student performance and engagement when they are not part of teacher-led classrooms.
What Happened
In an analysis published by Khan Academy, Chief Learning Officer Dr. Kristen DiCerbo argues that the sudden boom in generative AI has flooded schools with products promising to fix low reading scores and mental health issues. History shows that schools often launch scattered technology pilots that lead to program fatigue and little academic progress. To break this cycle, DiCerbo recommends that school leaders evaluate every proposed tool using four criteria: importance, viability, governability, and the existing evidence base. By focusing strictly on learning outcomes, which schools can control, educators can avoid adopting technology simply because it is new. As we previously reported, deliberate structural goals are critical for student growth.
The Bigger Picture
Academic research supports the warning against unstructured AI. A 2026 randomized controlled trial published on SSRN evaluated a course-integrated generative AI tutor at a large public university. Students with access to the AI tutor saw their final grades drop by 0.37 standard deviations. They also had a 0.90 standard deviation decrease in learning management system engagement. The drops were even worse for first-generation college students, which suggests that poorly integrated AI displaces effective study habits.
This failure is often structural. According to a study on the human readiness gap, school districts frequently focus on technical procurement, such as funding and device counts, while ignoring the training teachers need to succeed. When districts run short-term tests, they often use highly enthusiastic volunteers. This creates distorted results that do not translate to ordinary classrooms, as detailed in an analysis of why school pilots fail.
Technology works when it is structured and led by teachers. A systematic review published in MATHEdunesa found that math software boosts student understanding, but only when tied to teacher-led lesson plans. Similarly, a study in Scientific Reports showed that students using adaptive AI tutoring tools like Edmentum Exact Path outperformed peers in achievement and long-term retention when the tool supplemented, rather than replaced, human instruction. Research on pre-practice instructional support shows that students make fewer mistakes when digital platforms explain rules and concepts before asking them to solve exercises.
What This Means for Families
For parents and educators, personalized AI tools are not a magic bullet. If a school introduces an AI tool as a standalone study aid without a structured curriculum, it can distract students and lower their grades. Flashy apps are also a risk. As we have previously covered, superficial rewards can undermine genuine motivation. Real progress happens when technology is paired directly with what students are already learning in class under a teacher's guidance.
What You Can Do
First, ask administrators if your school's AI or EdTech program is used as a voluntary add-on or if it is integrated into the daily, teacher-led curriculum.
Second, check that the digital learning tools your children use provide explicit rules and explanations before prompting students to answer questions, rather than just acting as digital workbooks.
Finally, push school boards to invest as much money and time into teacher training as they do into software licensing.