AI can save teachers time. But is it making teaching better?

That distinction is at the heart of From Efficiency to Efficacy: Measuring AI's Instructional Impact, a new white paper written by my colleague Erica Price Burns and commissioned by Brisk Teaching. It's out today, with a foreword from Digital Promise CEO Jean-Claude Brizard. 

The paper starts with a question districts are wrestling with as AI tools settle into classrooms: What should we actually be measuring?

The default answer so far has been time saved, and for understandable reasons. Time is easy to count. A study of 542 educators across four Wisconsin districts found that most teachers saved meaningful time using AI, and nearly half reinvested it in instructional work with students.

But time saved is a starting point, not a finish line. The more useful measure is whether AI increases what Erica calls high-quality instructional moments—practices that learning science has already validated—and delivers them more often and more consistently. The paper pulls from decades of research that have helped define what a high-quality instructional moment looks like – timely feedback, targeted instruction and tutoring support, differentiated content for diverse learners. 

The gap isn’t a lack of research or knowledge. It's the teacher's capacity. Dave Hinrichs, who leads digital platforms and pedagogy at Johnston County Public Schools, puts the feedback problem plainly: It's "rarely timely and specific at the same time, and almost never actionable." The paper argues that AI can help close that gap and that districts should hold vendors accountable for demonstrating this, using measures weighted by instructional value rather than raw usage. As Sandra Rose of Prince George's County Public Schools says, "Usage is one thing. How you're using it is more important."

That question runs through today's links, too. In EdSurge, a teacher describes using AI to build AI-resistant assignments, while students tell the outlet they want a say in how AI policies are written. Tech & Learning looks at designing student work that demonstrates real thinking in the age of AI. Different angles, same underlying challenge: The tools are here, and the field is still determining what good use looks like.

Rose shares a story in the paper that has stayed with me. During a teacher shortage, she taught U.S. government to a class of more than 50 students, spending the semester trying to differentiate instruction while making every student feel seen. "I would have killed for something like this," she says of today's tools. "I often think about what I could have done."

That's the possibility districts should be measuring.

Looking for your next opportunity in education? W/A Jobs features 3,100+ career opportunities from 300+ organizations across the education industry.