Most of the AI-in-education conversation still hovers around higher education and secondary classrooms. Policy papers, assessment frameworks, GenAI guidelines, all of it assumes we are talking about students who are already producing essays and research papers. But what about the youngest learners? What does AI education look like in a second-grade classroom where half the students are still learning to read?
ISTE and General Motors published Hands-On AI Projects for the Classroom: A Guide for Elementary Teachers to answer that question. Written by Nancye Blair Black and Susan Brooks-Young, the guide offers four complete projects for K-5 teachers, each one built around a core AI concept: what AI does well (and what it doesn’t), training data and machine learning, sensors and perception, and navigation. The projects are student-driven, standards-aligned, and designed for teachers who may have zero technical background in AI.
I am covering this guide because the pedagogy is strong. The AI tools it references are not.
Black and Brooks-Young take on a common assumption early: that K-5 students are too young for AI concepts. They argue that “students in these grade levels have been taught skills that lead to an early understanding of how AI works: pattern recognition, sequencing, categorization, sorting, navigation skills, map reading, and even knowledge of animal senses” (p. 6-7). The infrastructure for AI understanding is already there. Teachers just need to name it.
I agree. And I would go further. Su, Ng, and Chu (2023) made a similar case in their review of AI literacy in early childhood education, finding that young children can grasp AI concepts when they are introduced through age-appropriate, hands-on activities. The ISTE guide puts that principle into practice. It has students sort picture cards, play AI guessing games, and design their own robots. The learning is physical before it is digital.
The authors also acknowledge a harder truth: “when it comes to the educators who are themselves in the early stages of learning about AI, very little is available to help them transfer what they are learning into meaningful, student-driven classroom activities” (p. 10). That sentence was written before ChatGPT existed. The teacher readiness gap it describes has only widened since. The RAND survey of U.S. teachers (Diliberti et al., 2024) confirmed that most K-12 educators are still figuring out where AI fits in their practice, and elementary teachers often get the least support.
What the Four AI Projects for Elementary Students Actually Teach
The guide’s four projects each address a different facet of AI, and they follow a consistent three-phase structure: Getting Started activities that hook interest and activate prior knowledge, Take a Closer Look activities that scaffold AI concepts through subject-area content, and Culminating Performances where students synthesize and reflect.
Project 1 asks students to identify tasks AI does well, like image recognition and game-playing, and tasks it doesn’t, like discerning emotions or making ethical decisions. Students play AI Tic-Tac-Toe, explore a bird-sound classification tool, and test AI guessing games like Akinator. The core lesson is clear: AI is good at narrow, specific tasks, not everything.
Project 2 is the strongest. Black and Brooks-Young walk students through the full pipeline of building a dataset: collecting data, labeling it, checking for bias, removing incorrect entries, and creating classification rules. The activity uses physical vocabulary picture cards. Fifty cards representing a category like “transportation,” with five cards deliberately misplaced. Students have to catch the errors, add missing categories, and write rules for classifying new items.
That is bias education for elementary students. Not a lecture about algorithmic fairness, but a hands-on experience of discovering that when a dataset is missing categories, the data becomes unreliable. Students who go through this activity understand, at a gut level, why training data quality matters. Most adults cannot say the same.
Project 3 draws a parallel between animal senses and the sensors AI robots use to perceive the world. Project 4 connects map reading and route planning to AI navigation systems. Both projects end with reflection questions that push students to consider societal implications: Should robots do tasks that humans currently do? Is it good to rely on navigation apps for every trip?
These are the right questions to ask eight-year-olds. They are the same questions the field is asking everyone else.
Where the Guide Shows Its Age
This is a pre-ChatGPT resource. The text generation tool it references is Write with Transformer, which runs on GPT-2. The AI examples are Siri, self-driving cars, and Netflix recommendations. There is no mention of generative AI or large language models.
That matters. The AI concepts the guide teaches, perception, representation, learning, natural interaction, societal impact, are timeless. The AI4K12 Five Big Ideas framework that structures these projects has not lost relevance. But the tools students actually encounter in 2026 are fundamentally different from the ones described here. A five-year-old today is more likely to have talked to ChatGPT than to have played AI Tic-Tac-Toe.
The guide also predates the explosion of AI literacy frameworks. UNESCO’s AI Competency Framework for Students (2024) now provides a progression model that maps AI literacy across age groups and levels. Chee, Ahn, and Lee (2025) developed an AI competency framework with variation by learner group. The ISTE guide’s alignment to its own standards and the AI4K12 framework was forward-thinking at the time, but teachers using it today would benefit from mapping its activities to these newer frameworks too.
I am not saying this to dismiss the guide. The pedagogical structure, unplugged activities paired with digital ones, scaffolded across three phases, grounded in student agency, is exactly right. Black and Brooks-Young built something that works. It just needs updating to reflect the AI tools and concepts students now encounter daily.
A teacher picking up this guide in 2026 should keep the structure, keep the unplugged activities, keep the reflection questions, and swap in current AI tools for the dated ones. The framework holds. The specific examples need refreshing to account for generative AI, conversational agents, and the fact that most students have already used these tools outside of school.
The pedagogy outlasts the technology. That is the highest compliment I can pay a curriculum guide.
References
- Black, N. B., & Brooks-Young, S. (n.d.). Hands-on AI projects for the classroom: A guide for elementary teachers. International Society for Technology in Education (ISTE). https://iste.org/ai
- Chee, H., Ahn, S., & Lee, J. (2025). A competency framework for AI literacy: Variations by different learner groups and an implied learning pathway. British Journal of Educational Technology, 56, 2146-2182. https://doi.org/10.1111/bjet.13556
- Diliberti, M. K., Schwartz, H. L., Doan, S., Shapiro, A., Rainey, L. R., & Lake, R. J. (2024). Using artificial intelligence tools in K–12 classrooms. RAND Corporation. https://www.rand.org/t/RRA956-21
- Su, J., Ng, D. T. K., & Chu, S. K. W. (2023). Artificial intelligence (AI) literacy in early childhood education: The challenges and opportunities. Computers and Education: Artificial Intelligence, 4, 100124. https://doi.org/10.1016/j.caeai.2023.100124
- UNESCO. (2024). AI competency framework for students. United Nations Educational, Scientific and Cultural Organization. https://doi.org/10.54675/JKJB9835