AI Can Save Educators Time. Here's What It Shouldn't Take Away.

New research suggests that the value of AI should be measured not only by the work it eliminates, but also by the instructional insight and human connection it preserves.

There is no question that educators need more time.

Course design, assessment, feedback, student support, and the many responsibilities surrounding teaching all compete for it. Against that reality, the promise of artificial intelligence is compelling: AI can help educators work more efficiently and redirect their limited time toward students.

But there is an assumption hidden inside many of our conversations about efficiency. We often treat time-consuming work as though its only cost is the time required to complete it. In teaching, however, the work frequently serves more than one purpose.

Reading a student submission may lead to a grade and a set of comments, but it also helps an educator recognize a misconception, understand how students are experiencing the course, and see where an assessment or instructional approach may need to change. If we automate the task without understanding everything that happens within it, we may save time while losing something equally important.

Educators Want AI that Fits the Work

This tension emerged in aresearch-practice partnership conducted by Digital Promise with support from D2L. The study examined how higher education faculty, staff, and instructional leaders were using D2L Lumi, a collection of AI capabilities embedded within D2L Brightspace.

The findings did not offer an uncomplicated story of either enthusiasm or resistance. Instead, they revealed a group of educators who were generally optimistic about AI but discerning about where and how it should enter their work.

Participants responded most positively to practical, bounded applications. Lumi Quiz, which generates assessment questions from course content, emerged as the strongest early use case. Educators valued its speed and straightforwardness, but they also valued that its questions could be inspected, edited, and approved before being added to a course.

That combination matters. The tool reduced effort without asking educators to relinquish control over the final instructional decision.

Participants also imagined AI taking on more of the routine work surrounding teaching. Some wanted it to function more like an administrative assistant than a content creator. They described possible support with announcements, calendar reminders, routine outreach, and personalized messages to students who had missed assignments.

These were not educators looking for AI to do the teaching for them. They wanted help clearing away busywork so they could devote more of their attention to the creative, relational, and student-facing dimensions of teaching.

Feedback Reveals a More Difficult Question

AI-supported feedback exposed a more complicated boundary.

The findings related to Lumi Feedback came from one participating institution, so they should be interpreted as an early signal rather than a broad conclusion. Even so, the experiences educators described raise important questions for the field.

Participants saw potential value in feedback that was detailed, consistent, and produced more quickly than an instructor might be able to provide it. They also encountered significant limitations. Some feedback was overly positive or sounded noticeably AI-generated. In other cases, the system introduced information that was not present in shorter student submissions. Educators also wanted more control over the kind of feedback the system generated.

As a result, instructors sometimes spent considerable time checking and revising the output. That raised an immediate efficiency question: If educators must inspect every comment closely, is the tool actually saving time?

But participants raised a deeper pedagogical question as well. Working through student submissions and developing feedback helped them understand what students found difficult, where conceptual misunderstandings were emerging, and how learners were experiencing the course. Providing feedback was not simply an output to be produced. It was part of how educators learned from and about their students.

This is the work inside the work, and it can become invisible when efficiency is our primary frame.

Saving Time Is the Mechanism, Not the Outcome

The Digital Promise research offers a useful starting point for evaluating these choices. Participants and researchers identified three criteria for assessing the success of instructor-facing AI: efficiency, effectiveness, and accessibility.

Efficiency asks whether AI reduces educator time and effort, including the time required to revise its output.

Effectiveness asks whether the output reflects what the educator intended, aligns with learning objectives, is actually used, and contributes to student engagement and outcomes.

Accessibility asks whether the technology and its outputs meet accessibility expectations and support principles such as universal design for learning.

These were proposed evaluation criteria, not confirmed outcomes of the study. That distinction is important. Access to AI is not the same as adoption, and adoption is not evidence of educational impact.

For institutional leaders, the framework also presents an opportunity to ask better questions before expanding an AI-supported practice:

  • What problem of practice are we trying to solve?
  • Is this work primarily repetitive, or does completing it produce meaningful insight into learning?
  • Can educators inspect, shape, and confidently stand behind the AI's output?
  • What evidence would demonstrate effectiveness beyond the number of times the tool is used?
  • How will we know whether the time saved is improving feedback, course design, student support, or learning?

The goal should not be to preserve every existing teaching task simply because it is familiar. Educators face genuine workload pressures, and AI can help reduce instructional friction. But we should be equally careful not to automate away the moments that allow educators to notice, interpret, respond, and care.

The most meaningful use of AI will not be the one that removes educators from the work. It will be the one that creates greater capacity while keeping them connected to student thinking.

Saving time matters. What educators are able to see and do with that time matters more.

Featured