Artificial intelligence (AI) is rapidly becoming part of the educational landscape. New tools promise to reduce teacher workload, personalize learning, and open new possibilities for instruction. But as AI adoption accelerates, one question matters more than the promise of the technology itself: How are educators actually experiencing these tools in practice?

A recent Teach For America survey of more than 3,400 first- and second-year educators across the country offers a revealing snapshot. The findings suggest that AI has achieved broad awareness and adoption among early-career educators, but the field is still grappling with a deeper question: how AI can best support student learning while honoring educators' professional judgment, values, and responsibilities. Approximately three-quarters of surveyed educators report using AI at least occasionally, yet relatively few consider it essential to their practice, highlighting a profession that is actively experimenting while still determining where these tools fit.

The role of AI in teaching and learning remains far from settled. As educators who are simultaneously learning their craft and navigating the rapid emergence of AI, new teachers offer a unique perspective on how the technology is reshaping classroom practice. Their responses reveal both enthusiasm for AI's potential and ongoing questions about how to use it in ways that strengthen student learning while maintaining educators' professional judgment, values, and responsibilities.

Three out of four new TFA educators surveyed report using AI at least occasionally. However, only 8.5% describe AI as essential to their teaching practice. Most educators fall somewhere in the middle: experimenting with AI, using it for selected tasks, and determining where it adds value within their classrooms and schools.

This pattern reflects an important reality about innovation: Adoption often precedes transformation. Teachers commonly begin by using new technologies to support existing workflows before fundamentally redesigning instruction around them.¹

Today's educators appear to be in that exploratory phase. AI has moved beyond curiosity, but many teachers appear to be still determining how these tools fit into the complex work of teaching and learning.

Among TFA educators surveyed, the most commonly reported benefit of AI is saving time. More than 60% of respondents report that AI helps reduce workload, while nearly half say it lowers stress or cognitive load. By comparison, fewer than 20% report direct benefits to student learning.

This may initially seem surprising given the ambitious claims often made about AI's educational promise. Yet it is consistent with other surveys on teacher AI use.² Our survey data are telling us that TFA educators most frequently report using AI for lesson planning, adapting instructional materials, and completing administrative tasks. At this point, student-facing applications remain less common. This does not necessarily mean AI lacks educational value. Rather, reducing administrative burden may be one of AI's most immediate contributions to education, allowing teachers to devote more time to planning, feedback, relationship-building, and other high-impact aspects of their work. As educators become more familiar with these tools and develop a stronger understanding of how AI can support learning, its value may extend beyond efficiency gains alone. Over time, AI may not only help teachers reclaim time, but also support instructional practices that accelerate student learning, provided its use is grounded in strong pedagogy, appropriate safeguards, and evidence of effectiveness.

These distinctions between AI use for efficiency versus learning mirror a broader conversation emerging across the field. A recent Brookings Institution report argues that the central question is no longer whether schools will use AI, but how they will use it.³ Drawing on a global study of educators, students, families, and researchers, Brookings concludes that AI can either enrich or diminish learning depending on its implementation. The report warns that overreliance on AI may undermine students' cognitive development, relationships, and agency, while well-designed uses of AI can enhance learning when embedded within strong pedagogical practice. This distinction helps explain an important finding from our data: while new teachers clearly see AI's value for improving efficiency, fewer report direct benefits for student learning. There is opportunity to identify the conditions under which AI meaningfully improves teaching and learning and to ensure there is strong professional development to support the appropriate applications of AI.⁴

Emerging large-scale studies show that the relationship between AI use and student outcomes depends on how the technology is used. In one recent study researchers found that AI-supported problem solving was associated with faster task completion but, in some contexts, weaker ability of students to retain what they learned.⁵ These findings reinforce the need for instructional approaches that leverage AI to support thinking rather than replace it and to connect AI use to evidence of knowledge retention, not simply evidence of efficiency.

What we're also hearing from TFA educators is that their concerns about AI appear to be driven less by technical barriers and more by ethical considerations.

More than half of TFA teachers expressed concerns about bias, overreliance, or broader societal consequences, while relatively few cited lack of training as a barrier.

These findings reflect a broader shift occurring across education. Increasingly, conversations about AI are moving beyond functionality toward questions of governance, privacy, ethics, and responsible use. Educators are asking important questions about how, when, and under what conditions AI should be used.

This perspective aligns with emerging scholarship arguing that education has historically approached new technologies with a default optimism, often treating adoption as inevitable. Some researchers contend that, particularly for rapidly evolving technologies such as generative AI, caution is not opposition but a legitimate and responsible response.⁶ They argue that because the potential benefits and harms of AI are still being understood, educators and institutions should be willing to critically evaluate new tools, establish safeguards, and, when necessary, limit or refuse uses that do not clearly align with educational goals.

Similarly, many district leaders emphasize the importance of implementing appropriate governance structures, privacy protections, and data-management practices before large-scale adoption. In some of the nation's largest school systems, the absence of clear AI policies has already generated significant pushback from parents, students, educators, and community members, prompting calls for greater transparency, stronger safeguards, and more opportunities for public input into AI decision-making.⁷ New York City, for example, recently delayed releasing districtwide AI guidance after receiving thousands of public comments and facing criticism from families and policymakers, while students themselves have increasingly stepped into the policy conversation, arguing for stronger guardrails around privacy, academic integrity, and the role of AI in schools.⁸

These conversations reflect a desire to ensure that innovation is accompanied by clear safeguards and thoughtful implementation, and that teachers are protected from unwittingly violating student privacy laws.

The Brookings Institution's Prosper, Prepare, Protect framework points in a similar direction.⁹ It calls on education systems to build AI literacy and capacity while simultaneously protecting student privacy, well-being, and cognitive development. Together, these perspectives suggest that the future of AI in education will depend not only on what the technology can do, but also on the judgment educators exercise in deciding when, where, and how it should be used.

Across the education sector, organizations are beginning to grapple with what educators need in order to navigate these concerns.¹⁰ New professional learning initiatives are focusing not only on technical proficiency, but also on helping educators develop judgment, discernment, and responsible AI practices.

At Teach For America, we believe educators should have a leading role in shaping how AI is used in teaching and learning. That requires more than technical proficiency. It requires the judgment to determine when AI can strengthen learning, when it cannot, and how to use it responsibly. This belief is reflected in our approach to educator development from day one. For the past four years, all incoming corps members have engaged in AI training as part of their pre-service experience. This year’s course, AI Fluency for pK–12 Educators, codesigned with Anthropic, intends to help educators navigate exactly the questions surfaced by this survey: how to use AI ethically, responsibly, and in ways that support rather than substitute for student learning. The course focuses not only on practical applications of AI, but also on developing the discernment needed to evaluate outputs, protect student privacy, understand limitations, and make informed decisions about when AI should and should not be used.

And as our educators gain years and experience in the classroom, TFA offers a continuum of training to learn about AI in the classroom, deepen their understanding through hands-on experimentation, and ultimately gain the dexterity to decide when AI is and is not appropriate for their learning objectives.

Those trainings align closely with what our teachers themselves are telling us. They are not asking for AI for AI's sake. They are looking for ways to use these tools responsibly, thoughtfully, and in service of students.

The future of AI in education will likely depend less on how quickly schools adopt new tools and more on whether educators can learn to use them in ways that strengthen teaching, protect student trust, and create meaningful learning opportunities. The most important work ahead may not be teaching educators to use AI, but helping the field determine how AI can best serve students.

The good news is that educators appear ready for that conversation. And if the findings from this survey are any indication, they are eager to approach it with both curiosity and care.​

1. Terada, Y. (2020, May 4). A powerful model for understanding good tech integration. Edutopia. https://www.edutopia.org/article/powerful-model-understanding-good-tech…
2. Langreo, L. (2026, January 12). More teachers are using AI in their classrooms. Here's why. Education Week. https://www.edweek.org/technology/more-teachers-are-using-ai-in-their-c…
3. Burns, M., Winthrop, R., Luther, N., Venetis, E., & Karim, R. (2026, January 14). A new direction for students in an AI world: Prosper, prepare, protect. Brookings Institution. https://www.brookings.edu/articles/a-new-direction-for-students-in-an-a…
4. Tan, X., Cheng, G. K. S., & Ling, M. H. (2025). Artificial intelligence in teaching and teacher professional development: A systematic review. Computers and Education: Artificial Intelligence, 8, Article 100355.
5. Rismanchian, S., Uzun, H., Matayoshi, J., Cosyn, E., & Kurd-Misto, E. (2026). Faster completion, less learning: Generative AI reduced study time on math problems and the knowledge they build [Preprint] (arXiv:2605.21629). https://doi.org/10.48550/arXiv.2605.21629
6. Shah, N., Stroupe, D., Yadav, A., Bernhard, T., Logan, C., Dyer, E. B., He, P., Krist, C. (S.), & Li, T. (2026). Refusing educational technology: Artificial intelligence, inequity, and the problem of critical optimism. Educational Researcher. Advance online publication. https://doi.org/10.3102/0013189X261466234
7. Turner, C. (2026, July 30). Adults have struggled to set rules for AI in school. These teens figured it out. NPR. https://www.npr.org/2026/07/30/nx-s1-5853571/students-set-ai-policy; Insignares, G. (2026, August 4). Los Angeles educators, parents want more of a say in AI's complicated future in classrooms. CBS Los Angeles. https://www.cbsnews.com/losangeles/news/los-angeles-educators-parents-w…
8. Zimmerman, A. (2026, June 24). NYC promised final school AI guidance by June. Now officials are hitting pause. Chalkbeat. https://www.chalkbeat.org/newyork/2026/06/24/nyc-education-department-d…
9. Burns, M., Winthrop, R., Luther, N., Venetis, E., & Karim, R. (2026, January 14). A new direction for students in an AI world: Prosper, prepare, protect. Brookings Institution. https://www.brookings.edu/articles/a-new-direction-for-students-in-an-a…
10. See for example: https://www.air.org/project/teaching-ai-testing-promise-and-empowering-…