A new white paper from the Center on Reinventing Public Education (CRPE) names our current reality: no shared measures of success, governance not keeping pace with the technology, and a market increasingly shaped by the priorities of ed-tech vendors. Generative AI carries real risk—”wicked problems” like data and identity theft, cognitive offloading, and threats to students’ future employment. But it also holds “wicked opportunities” to solve problems that have plagued public education for decades: disengaged students, ineffective special education programs, burnt-out teachers, and high schools disconnected from real career pathways.
Four gaps are standing in the way:
- The vision gap. There is no shared or prevalent definition of what AI should accomplish for students across school systems, states, or the broader education field. Districts are buying what neighboring districts bought, not what fits their own goals.
- The coherence gap. AI tools often arrive in classrooms as standalone additions to an already fragmented instructional environment. Schools serving students with the greatest needs are left with the least capacity to make sense of it all.
- The quality gap. The evidence base for AI in education remains thin, fragmented, and poorly aligned with the decisions educators and policymakers actually need to make.
- The leadership and policy gap. Even district and state leaders with a compelling vision have few roadmaps for turning it into coherent organizational change.
These four gaps converge on one structural problem: AI is being layered onto existing systems rather than used to redesign them. Left unchecked, the risks are real—wider gaps between well-resourced and under-resourced schools, more entrenched factory-model instruction, and outsized power in the hands of AI companies who answer to shareholders, not students. But with the right research investments now, education leaders can shift that trajectory instead of just muddling through it.
The report outlines three priorities for research:
- Measurement and Policy — the evidence, governance structures, and policy conditions needed for responsible innovation.
- Coherence and Scale — how educators, schools, and districts can use AI to actually support teaching, learning, and student success.
- Whole School Redesign — how AI can enable fundamentally different models of learning, teaching, and schooling, not just faster versions of the same ones.
Within each, the authors pose the urgent, high-stakes questions researchers, funders, and districts should be racing to answer now—from what counts as student success in the age of AI, to how AI can transform IEP development from a compliance exercise into genuine individualized support, to what new staffing models could free teachers to focus on the relationships students are starving for. This white paper calls on researchers to get embedded in this work now, funders to require evidence generation as a condition of investment, and system leaders to treat “what works” as an open empirical question rather than a vendor claim. The market will not wait for the field to catch up.
