With rapid advances in generative artificial intelligence, teachers and school leaders are looking for answers to complex questions that will help them successfully incorporate technology into their classrooms while ensuring that students actually learn what they are trying to teach them.
Justin Reich, an associate professor in MIT’s Comparative Media Studies and Writing Program, hopes a new guidebook published by the MIT Education Systems Institute can help K-12 educators decide what AI policies and guidelines to create.
“Throughout my career, I have strived to be someone who studies education and technology and communicates that research to the people working in the field,” Reich says. “When something difficult happens, I try to jump in and help.”
Published this fall, “A Guide to AI in Schools: A Perspective for the Confused” was developed with the support of an expert advisory board and other researchers. The project includes input from more than 100 students and teachers across the United States who share their experiences teaching and learning using new generative AI tools.
“We try to advocate a spirit of humility when considering AI in schools,” Reich says. “We’re sharing some examples from educators about how they’re leveraging AI in interesting ways. Some of them may turn out to be robust, some of them may turn out to be flawed. And we won’t know which ones for a long time.”
Find answers to your AI and education questions
This guidebook is intended to help K-12 educators, students, school leaders, policy makers, and others gather and share information, experiences, and resources. With the advent of AI, schools are faced with multiple challenges, including how to ensure academic integrity and maintain data privacy.
Reich cautions that the guidebook is not prescriptive or definitive, but is meant to help provoke thought and discussion.
“Writing a guidebook to generative AI in schools in 2025 is a bit like writing a guidebook to aviation in 1905,” said the guidebook’s authors. “In 2025, no one will be able to say the best way to manage AI in schools.”
Schools are also struggling to measure what student learning loss looks like in the AI era. “What would actually happen if we used AI to avoid productive thinking?” Reich asks. “When you think that teachers are no longer providing content and context to support learning, and students are no longer doing the exercises that accommodate the content and provide context, that’s a serious problem.”
Reich invites those directly affected by AI to help develop solutions to the challenges posed by AI’s ubiquity. “It’s like observing conversations in the teacher lounge and engaging students, parents and others to participate in how teachers are thinking about AI, what they’re seeing in their classrooms, what they’re trying, and what’s going on,” he says.
In Reich’s view, the guidebook is ultimately a collection of hypotheses expressed in interviews with teachers, an informed first guess about the path schools may take in the coming years.
Create educator resources with podcasts
In addition to the guidebook, Teaching Systems Lab also recently produced “The Homework Machine,” a seven-part series of Teachlab podcasts that explores how AI is reshaping K-12 education.
Reich collaborated with journalist Jesse Dukes to produce the podcast. Each episode tackles a specific area and asks important questions about challenges related to the implementation of AI, poetry as a tool for student engagement, post-COVID learning loss, pedagogy, book bans, and other issues. Podcasts allow Reich to share timely education-related updates and collaborate with those interested in advancing research.
“The academic publishing cycle is not well-suited to helping people with the kind of short-term challenges that AI poses,” Reich says. “Peer review takes a long time, and the research produced is not always in a format that is useful to educators.” Schools and districts are working on AI in real time, bypassing proven quality control measures.
Podcasts can help reduce the time it takes to share, test, and evaluate AI-related solutions to new challenges, and can be useful for training and resource creation.
“We hope the podcast sparks thought and discussion and allows people to be informed by other people’s experiences,” Reich says.
The podcast was also produced as a one-hour radio special and broadcast on public radio stations across the country.
“We are groping in the dark.”
Reich offers a candid assessment of where we stand when it comes to AI and its impact on education. “We’re groping in the dark,” he says, recalling past attempts to quickly bring new technology into the classroom. These failures highlight the importance of patience and humility as AI research continues, Reich suggests. “AI has bypassed the normal procurement processes in education and just appeared on children’s phones,” he points out.
“We’ve been really wrong about technology in the past,” Reich says. For example, even though school districts are spending on tools like smart boards, research shows there is no evidence that they improve learning or outcomes. In the new article of the article conversationhe argues that teachers’ early guidance in areas such as web literacy created the bad advice that still exists in our education system. “We taught students and educators not to trust Wikipedia and to search for trustworthiness markers on websites, and both turned out to be wrong,” he recalls. Reich wants to avoid making similar judgments about AI, and recommends avoiding second-guessing instructional strategies that leverage AI.
These challenges, combined with the potential and observed impacts on students, significantly increase the risk for schools and students’ families in the AI race. “Educational technology always causes anxiety for teachers,” Reich points out. “However, the scope of AI-related concerns is much broader than other technology-related areas.”
The dawn of the AI era is different from how technology has been introduced into classrooms in the past, Reich said. AI was not adopted like other technologies. It just arrived. It is currently upending educational models and, in some cases, complicating efforts to improve student outcomes.
Reich was quick to point out that there are no clear, definitive answers to the effective implementation and use of AI in the classroom. Those answers don’t currently exist. Each resource Reich helped develop invited engagement from a targeted audience and aggregated valuable answers that others might find helpful.
“We can develop long-term solutions to schools’ AI challenges, but it will take time and effort,” he says. “AI is not like learning to tie knots. We don’t yet know what it is or what it will be.”
Reich also recommends learning about AI implementation from a variety of sources. “Pockets of decentralized learning help us test ideas, search for themes, and gather evidence about what works,” he says. “We need to know whether using AI actually improves learning.”
Although teachers do not have a choice in the presence of AI, Reich believes it is important to seek teacher input and involve students and other stakeholders to help develop solutions that improve learning and outcomes.
“Compete to be the right answer, not the first,” says Reich.
