The robots are here (and they have feedback): Can AI improve teacher effectiveness?

If there’s one constant in teaching, it’s change. Whether it’s new curricula, new professional learning, or new ways to organize the classroom, teachers are always adapting. But in recent years, no innovation has promised more transformation—or sparked more debate—than artificial intelligence (AI).

In fact, AI’s presence in education is growing rapidly. By Fall 2024, 43% of teachers had participated in at least one AI training session, a 50% increase from the previous spring.¹ Despite this surge, the Center for Reinventing Public Education highlights a key challenge: While district leaders see AI as a game-changer for easing teachers’ workloads—helping with lesson planning, grading, targeting student interventions, and countless other tasks—many teachers still have minimal experience with AI tools. For example, fewer than 40% of teachers report using the popular AI tool ChatGPT at work—and that number drops to just 30% among female teachers,² who make up the vast majority of the teaching workforce, especially in elementary schools.

This gap raises an important question: Can AI truly support teachers in ways that improve teacher effectiveness?

This month’s District Trendline explores the latest research on AI’s role in improving teacher effectiveness, including how it can be leveraged to provide coaching and feedback, support teacher preparation and workflow processes, and even strengthen large-scale educational research.

AI as a feedback and coaching tool

Whether they’re still in their preparation program or have been in the classroom for decades, teachers can benefit from personalized, targeted feedback. But school leaders have long to-do lists, so providing meaningful feedback—especially beyond required evaluations—often falls by the wayside. Research suggests this is a key area in which AI can offer valuable support.

A 2023 randomized controlled trial found that AI could be used to improve teachers’ “uptake of student ideas.” Researchers collected class recordings and data from more than 900 instructors and 10,000 students in an online course at Stanford University. They deployed an automated AI tool to analyze class transcripts for examples of instructor uptake, defined as any time a teacher “acknowledges, revoices, and uses students’ ideas as a resource.” For instance, if a student suggests that plants might grow faster when exposed to music, a teacher demonstrating uptake could integrate the comment into the lesson, perhaps by launching an experiment to test the hypothesis. Past research suggests that uptake is a critical component of strong instruction.³

Using the data collected, researchers developed AI-powered dashboards to provide feedback to teachers that included statistics like teacher talk time, examples of strong uptake, and tips for improvement. The study found that instructors who used these dashboards increased their uptake of student ideas by 13%. These findings highlight the potential of AI as a low-cost, high-impact intervention for improving classroom discourse and instructional quality. However, the study had a few elements that may make it hard to generalize to other settings. The instructors were volunteers rather than trained teachers, and they provided instruction in an online setting.

For more traditional “brick and mortar” classrooms, a separate 2024 study led by the same researcher revealed that automated feedback using AI can provide significant value to K–12 teachers in schools. The study involved over 200 Utah teachers who recorded their lessons using an app that transcribed their instruction and provided feedback on elements such as teacher talk time, wait time, and instructional techniques. While all participating teachers had access to the standard app, a treatment group received additional feedback from the platform and automated emails specifically targeting the teachers’ use of “focusing questions”—questions that encourage students to explain their thinking and reflect on their understanding. For instance, while “What is the slope of this line?” pushes students to identify a concrete answer, “What do you mean by the angle of the line?” is a focusing question that challenges students to reflect more deeply on their conceptual understanding.

Teachers who received feedback on using focusing questions increased their use of these questions by 20% compared to the control group. However, engagement with the feedback was low—about a third of teachers never opened their feedback emails, and fewer than a quarter interacted with the feedback in the app. Notably, while the feedback influenced question use, it had little effect on other aspects of instruction, such as student talk time or how often teachers incorporated student ideas.

Conversations with participating teachers revealed that while some educators found it valuable for self-reflection and refining their questioning strategies, others expressed concerns about the accuracy of the feedback, which made them less likely to engage with it.

What’s the upshot? While these types of AI tools are still evolving and may occasionally misrepresent or fail to accurately assess what is happening in the classroom, this study highlights the potential of such tools to provide a cost-effective approach to supporting teachers with timely feedback.

District Spotlight: Recognizing the power and potential of AI, EdNovate Charter Schools (CA) is piloting AI tools designed to enhance teacher effectiveness through feedback and coaching. However, given the emerging nature of this technology, the district has established clear guidelines to ensure responsible implementation, focusing on safety and privacy, equity, integrity, accountability, and flexibility. These measures help safeguard students and teachers while maximizing the benefits of AI.

AI to simulate student interactions in teacher prep

Another emerging use of AI to improve teacher effectiveness is AI-assisted simulated classrooms in teacher preparation. While simulated classrooms are not a new phenomenon, these simulations often involve trained actors or other individuals playing the role of students. This can be expensive and hard to scale. Emerging research shows that AI-simulated students can provide valuable learning opportunities for teachers, likely at a lower cost than simulations with trained actors. While this is not a substitute for rich clinical practice, it may provide additional practice opportunities that mimic real student scenarios.

A 2023 analysis examined the effects of AI-simulated students by allowing aspiring teachers to act as teaching assistants during “office hours” with two AI-generated learners. The study found that this approach gave teachers a safe and flexible environment to practice and iterate, enabling them to spend more time crafting thoughtful feedback, developing clearer examples, refining language for inclusivity, and exploring multiple instructional strategies to help students reach their learning goals. Additionally, because the teacher candidates could repeat the simulations (and therefore redo their interactions with the AI students), data from participants revealed that interacting with AI-simulated students provided valuable opportunities for iterative thinking, deeper content exploration, and real-time problem solving. This allowed teachers to strengthen their own understanding and adapt their teaching strategies even in situations where they did not immediately know how to respond. Notably, it was the opportunity to rehearse and adjust their teaching across sessions—not feedback from the AI—that contributed to these improvements.

AI to support workflow management

A small-scale longitudinal study of 24 U.S. public school teachers, all new to using AI, found that participants fell into three distinct categories based on their AI usage throughout the school year:

1. Those who used AI for both generating materials and managing their workflow (lesson planning, grading, emailing, etc.)
2. Those who used it solely for generating materials
3. Those who did not use it at all

Teachers in the first group—who leveraged AI for both workflow planning and content creation—reported greater productivity gains in terms of workload management and work quality. In this context, workflow planning refers to how teachers used AI to help them decide which tasks to prioritize or how to structure lessons more effectively. In contrast, teachers in the second group, who only used AI to generate materials, did not report the same benefits. The researchers suggest that productivity gains stem from AI’s ability to “nudge teachers toward new and better tasks they may not have done otherwise.” Meanwhile, relying solely on AI for output creation helped teachers feel like they “got it done” rather than “got it right.”

The study also found that teachers with more classroom experience used AI as a “planning partner” to refine and iterate on their instructional strategies, while less experienced teachers tended to submit more prompts, likely experimenting with AI but limiting any time savings.⁴

As AI tools become more embedded in the teaching process, it is important to recognize their role as a complement to, rather than a replacement for, vetted high-quality instructional materials (HQIMs). Unlike HQIMs, which undergo rigorous review and align with evidence-based practices and state standards, AI-generated content can vary in quality. Ensuring that AI serves as a support for thoughtful planning rather than substitute for trusted materials will be key to maximizing its benefits without compromising instructional integrity.

District Spotlight: St. Vrain Valley School District (CO) offers a yearlong professional development program focused on integrating AI into classroom instruction. For instance, music teachers have created accompaniment tracks using AI, while other teachers have used AI to develop discussion questions to enrich vocabulary learning experiences. The program includes monthly training sessions and professional learning networks, providing educators with ongoing support and collaboration opportunities. Additionally, the district utilizes an AI coaching platform that empowers teachers through self-observation, goal setting, and action planning, helping them refine their instructional practices and effectively leverage AI in their teaching.

AI to strengthen educational research

Beyond its impact on individual classrooms, emerging research highlights how AI can aid in addressing broader educational questions—such as how educators distribute their attention across students. By analyzing over 1 million educator utterances—individual segments of speech from virtual group tutoring sessions—researchers used AI to systematically measure which students receive the most attention and the types of attention they receive.

The study found that while educators generally prioritize lower-achieving students, disparities emerge across demographic lines. Girls received less attention than boys in mixed-gender pairs, even when they were the lower-achieving student. Similarly, low-achieving Black students received additional attention compared to their higher-achieving Black peers—but only when paired with another Black student. When paired with a non-Black peer, they did not receive additional attention. Additionally, higher-achieving English learners (ELs) received significantly more attention than their lower-achieving EL peers when grouped together, suggesting that tutors may find it easier to engage with EL students who have stronger literacy skills, potentially reinforcing disparities in instructional support.⁵

These findings highlight how AI can help conduct large-scale analyses, uncovering subtle yet meaningful patterns in educator-student interactions. By revealing once-hidden disparities and trends, AI not only helps inform more equitable and effective educational strategies, but also provides educators with valuable insights to identify and address potential biases in their teaching practices.

Conclusion

Ultimately, the recent surge in AI tools presents both exciting opportunities and significant challenges for improving teacher effectiveness. While early research highlights promising benefits—from AI-powered feedback to simulated student interactions—there are still critical unanswered questions about the scalability, accuracy, and long-term impact of these initiatives.

Despite these challenges, AI’s potential to support teachers and improve instruction is undeniable. However, maximizing its benefits will require strategic implementation, ongoing research, and thoughtful safeguards to ensure that students and educators, not algorithms, remain at the center of the learning process.