What concerns educators and students
There are clear themes in what users say most frustrates and concerns them.
Teachers are frustrated by products that let students off the hook by giving away the answer or that are easy to "game."
Teachers worry about students using AI in ways that limit effortful thinking and shortchange long-term learning. Teachers get frustrated when the loopholes in product design make it appear as if students made progress or mastered content when they actually haven't (e.g., students can refresh the question list and answer the same question five times rather than answering five different questions). We saw and heard about multiple examples of students using a general-purpose chatbot to answer questions posed by an AI-powered instructional tool. Teachers get deeply frustrated by this and rightly worry that this behavior limits long-term learning.
Teachers feel that AI-powered products work better for some students than others.
Across products, many teachers report that stronger, more independent readers and more self-motivated students benefit the most, whereas students with more significant learning needs require more support to benefit. On the flip side, many teachers also worry that the ceiling on most products is too low for the most advanced students.
Students and teachers get frustrated when AI feedback misreads student work or does not specifically and succinctly explain what students need to do differently to meet expectations.
Students and teachers have little patience for products that mark correct answers as incorrect because a student's response doesn't fit an expected format. Students and teachers don't like waiting for feedback. Students lose interest when a tool's feedback feels generic. Some students get particularly frustrated by feedback that appears to come from their teacher but that they believe was generated by AI, feeling like their teacher did not actually read their work. Though they want the feedback to be specific and actionable, students and teachers dislike it when the feedback is too long or detailed.
The leaders, teachers, and students (and product developers) we talked to worry about instruction that includes too much screen time.
There is wide agreement that instruction should be a balance of learning time on and off screens and that relationships between teachers and students, and students and their peers, are central to learning. Even when products offer value, leaders, teachers, students, and product developers feel that there is an upper limit of time that students should spend on any given product and on total screen time.
Teachers and students worry about how other teachers and students use products to take shortcuts.
Students report that their peers use general-purpose chatbots—whether permitted by their school or not—to complete assignments. Teachers worry that other teachers might use AI to get out of parts of the job that they believe are important (e.g., using AI to generate lesson plans, rather than thinking it through themselves). Though no one we interviewed self-identified as participating in these practices, we frequently heard concern that they are widespread and damaging to teaching and learning.