The online interactive magazine of the Association for the Advancement of Artificial Intelligence

by David Touretzky, Chrisina Gardner-McCune

There has recently been an explosion of activity in K-12 AI education. As the AI4K12 Initiative ( works to develop national guidelines for teaching AI in US schools, similar efforts have begun in Canada, the U.K., China, and elsewhere. New curricula are being piloted, new tools developed, and new AI electives launched in middle schools and high schools. There have been national and international workshops, and ongoing discussions online about teaching AI in K-12. This symposium brought together the growing community of AI educators, researchers, curriculum designers, and tool developers to discuss the current state of AI education in K-12.

The “Teaching AI in K-12” symposium was the successor to 2018’s inaugural “AI for K-12” symposium, and expanded from a one-day, 40 participant event to two days and 75 participants. To accommodate the schedules of working K-12 teachers, we departed from the standard format by beginning on a Friday and continuing all day Saturday. The symposium was organized by the AI4K12 Initiative ( The initiative is a joint project of AAAI and CSTA, the Computer Science Teachers Association, with funding from the National Science Foundation. A generous gift from the AI Journal provided travel scholarships to 11 K-12 teachers, while support from AAAI allowed all K-12 teachers to register at a reduced rate.

The symposium brought together university researchers doing AI curriculum development and AI education research, industry players (IBM, Microsoft, ReadyAI), nonprofits (ISTE, AI4ALL,, Actua, Global STEM Education Center), publishers (Tumblehome Books, Tinker Toddler Books), school administrators, and K-12 teachers who are currently teaching AI topics or planning to introduce them into their classes. Strong international interest in K-12 AI education was evidenced by presenters and attendees from Canada, the U.K., France, Brazil, and China, plus pre-recorded video presentations from Austria and Sweden.

The symposium began with a review of progress of the AI4K12 Initiative in developing national guidelines for teaching AI in US schools. The Initiative has released a list of “Five Big Ideas in AI” that will form the framework for these guidelines. (For details, see “A Year in K-12 AI Education” in the Winter 2019 issue of AI Magazine.) The remaining presentations were organized around several themes: Middle and High School AI Curricula; AI Teacher Professional Development; Pedagogy; Ethics; International Perspectives; Informal Learning; and AI Tools. In addition there was a panel discussion offering perspectives from five K-12 teachers who serve on the AI4K12 Working Group, an “AI Playground” with hands-on demonstrations of hardware and software for teaching AI, and an interactive learning activity where participants simulated learning in a 3 layer neural network. The program was rounded out by several 5 minute “lightning talks”, some live and some pre-recorded. Among these was a talk by a high school student (live) about her experience attending a machine learning summer camp, and a recorded talk by the author of a series of AI and machine learning books for toddlers whose video included a shot of her infant daughter being driven around by a Roomba.

Friday’s keynote talk by Hal Abelson of MIT, “From Computational Thinking to Computational Action”, included several inspiring examples of students building apps with significant impact on their communities, such as helping people find sources of clean water. Abelson is the creator of the App Inventor programming framework used in these projects. App Inventor is now being extended with AI primitives such as a trainable visual classifier.

Saturday’s keynote by Ning Wang of USC and James Lester of North Carolina State University was titled “Building a Research Foundation for K-12 AI Education”. While we’re still in the early stages of curriculum development and tool building, it’s not too early to begin asking how we will measure their effectiveness and develop a body of knowledge about effective techniques and best practices, as has been done in other subject areas. Their talk was immediately followed by another noteworthy talk, by Miles Berry of Roehampton University, who spoke on “Computing in English Schools: Lessons to Learn for AI Education”.

The take home lesson from the symposium is that K-12 AI education is developing rapidly, with multiple curriculum and professional development projects under way by large and small players worldwide, and adventurous teachers striking out on their own to find what works. The attendees were invigorated and eager to see what the next year will bring. Slides from the live presentations and links to the video presentations are available on the web site.

David Touretzky, Chrisina Gardner-McCune, Fred Martin, and Deborah Seehorn served as co-chairs of the symposium.

David Touretzky is a research professor in the Computer Science Department and the Neuroscience Institute at Carnegie Mellon University, and chair of the AI4K12 Initiative.

Christina Gardner-McCune  is a CS Education researcher and an assistant professor in the Computer & Information Science & Engineering Department at the University of Florida, and co-chair of the AI4K12 Initiative.