From Overwhelmed to Equipped: The Real Story of AI in Today's Classrooms

This article was written by our partners at Digital Promise.

When Digital Promise started implementing AI in classrooms more than 6 years ago, AI tools were being built for specific classroom applications, like getting automated feedback on student writing, and mostly being implemented by early adopters. We were beginning to learn how AI might support personalization or assessment, though those early adopters expressed concerns about the accuracy and the emotional impacts of these kinds of tech. 

Historically—as with advancements like computational thinking—such small successes and challenges early on have provided product developers, researchers, and educators with an evidence base that yields insights for best practices, professional learning, and a grassroots model of adoption. This helps provide a roadmap for educators navigating new technologies. 

Instead, in 2022, ChatGPT and other generic AI-enabled tools expanded rapidly.

Since that time, AI has entered classrooms so quickly that adoption has outpaced research into best practices. Educators have been forced to make quick, reactive decisions about policies and guidelines for safety and learning. For example, assignments that were previously rigorous can now be completed in seconds, but teachers haven’t had time or support to think about how to redesign instruction. 

As a result, some educators are outright banning the technology, leaving learners left without the skills and experience to thrive in a world where emerging technologies are rapidly integrating across industries. So, how can we find a middle ground? How can we equip learners with knowledge and skills to make informed decisions and sound judgments about when and how to leverage AI in and out of school?

AI Promises vs. Pedagogy

This question has become increasingly urgent as tech leaders have made big promises in recent years, such as that AI would lead a personalized learning revolution where each student has a chatbot helping them close gaps at a rate their teachers, with a classroom full of students, never could. Admin hoped that it might be able to provide more actionable data and insights. Simultaneously, teachers hoped that AI might help lighten their load and support personalized learning. 

Yet, it probably comes as no surprise that chatbots have not fulfilled their promise as tutors. Why?

• As you well know, teaching is far from simple. Many LLMs (Large Language Models) don’t have the pedagogical underpinnings to effectively scaffold the learning process in the way an effective teacher can. For example, a teacher can attune their instruction to a student’s zone of proximal development to ensure students struggle enough to deepen their learning (but not so much that they disengage), a process which LLMs cannot replicate.
  

• AI might pose a risk to your students’ thinking. AI can lead to cognitive offloading, which can be helpful if that load is extraneous, but disastrous if that load is essential to a student’s learning. Because LLMs are prone to hallucinations, those who use AI need strong background knowledge and critical thinking skills to assess its output.
  

• AI might not help students in the long term. In some instances, student outcomes improve with the use of generative AI tools, but those gains disappear when the tool does, casting doubt on whether AI can actually lead to a meaningful transfer of learning. 

How AI is Actually Working in Education

However, if we support students in self-regulated learning, they can make informed decisions about how to use AI for learning to enhance their uniquely human skills. 

• In situations where students recognize extra cognitive load, they can decide how to use AI to remove barriers to their learning. For example, a student with executive functioning challenges might use AI to generate a plan for completing an assignment over the course of several weeks. 
  

• Where it is helpful, AI can be a tool that is specifically designed for learning. Instead of using foundational models directly, specific learning tools that use AI can support students’ durable skills (such as collaboration) by supporting rich in-class discussion. 
  

• AI can help you support your learners. For teachers, AI can save up to 6 hours a week by reducing administrative tasks or modifying materials to meet students’ needs. This extra time can be dedicated to providing individual support or mentorship —creating a sense of belonging that is key to student success. 

All of these point us back to the need for both teachers and learners to build AI literacy, so that when they start to incorporate the use of AI, they make strong, informed choices that enhance—rather than detract—from learning. AI literacy is the knowledge and skills that enable humans to critically understand, evaluate, and use AI tools to safely and ethically participate in a digital world. 

Building Our AI Literacy in the Classroom

To build AI literacy, students need explicit instruction in what AI is, how it works behind the scenes, and when it’s appropriate to use. But there is a dearth of easy-to-understand resources, especially for elementary students and teachers. Where is one to start?

That’s why Digital Promise partnered with BrainPOP to develop their new AI Literacy collection, coming Back to School 2026. This collection of animated movies, featuring Moby and friends, helps answer the biggest questions students (and teachers!) have about AI—equipping them with the knowledge and skills to navigate this technology critically and ethically.

Struggling with where to get started? Join us for our webinar “What Does AI Literacy Actually Look Like in Classrooms?” July 15, 2026, to learn actionable strategies for how teachers are developing foundational knowledge about AI in themselves and their students. 

---

Megan Pattenhouse is a learning designer, researcher, and former elementary educator with over a decade of experience helping teachers implement research-aligned technology at organizations like Digital Promise, Newsela, and Amira. She holds an MS in Digital Age Learning from Johns Hopkins and is a firm believer in the power of inquiry-based, student-centered instruction. 

Dr. Kelly Mills is a researcher and AI Literacy expert with a focus on bridging the gap between research and practice. As the Senior Director of Research for the Powerful Learning team at Digital Promise, she leads initiatives that integrate emerging technologies into K-12 environments, most notably by creating meaningful opportunities for students to use, understand and evaluate AI. She holds a Ph.D. in Curriculum and Instruction from the University of Maryland, College Park.

Additional thanks to Digital Promise contributors Dr. Judi Fusco, Dr. Chris Wegemer, Dr. Quinn Burke, Dr. Jeffrey Starr, Martika Parkinson, and Nancy Chou.