Twenty Constructionist Things to Do with Artificial Intelligence and Machine Learning - Paper overview

Aug 5, 2026

Upload by Liam

This paper proposes twenty constructionist activities using artificial intelligence and machine learning, building on Seymour Papert and Cynthia Solomon’s 1971 memo “Twenty Things to Do With a Computer.” Drawing from both inherited and new ideas across science, mathematics, and the arts, it emphasizes children’s engagement and developing insight into their own cognitive processes. The work argues for designing personally relevant AI/ML applications rather than relying on isolated models or off-the-shelf datasets. It also highlights the social dimensions of data production and the need to address harmful algorithmic biases and consequences.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages12

Reading Time30 min

AccessFree to download

Abstract

Background: Papert and Solomon’s “Twenty Things to Do With a Computer”

Why AI/ML Education Matters

From artificial-teachers to learner-created projects

Constructionist AI and Education Framework

Roles and approaches for constructionist AI in learning

Cite as: Kafai, Y. B. & Morales-Navarro, L. (2023) . Twenty Constructionist Things to Do with Artificial Intelligence and Machine Learning. In Proceedings ofFablearn/Constructionism 2023. Twenty Constructionist Things to Do with Artificial Intelligence and Machine Learning Yasmin Kafai & Luis Morales-Navarro, University of Pennsylvania ABSTRACT In this paper, we build on the 1971 memo “Twenty Things to Do With a Computer” by Seymour Papert and Cynthia Solomon and propose twenty constructionist things to do with artificial intelligence and machine learning. Several proposals build on ideas developed in the original memo while others are new and address topics in science, mathematics, and the arts. In reviewing the big themes, we notice a renewed interest in children’s engagement not just for technical proficiency but also to cultivate a deeper understanding of their own cognitive processes. Furthermore, the ideas stress the importance of designing personally relevant AI/ML applications, moving beyond isolated models and off-the-shelf datasets disconnected from their interests. We also acknowledge the social aspects of data production involved in making AI/ML applications. Finally, we highlight the critical dimensions necessary to address potential harmful algorithmic biases and consequences of AI/ML applications. KEYWORDS constructionism, machine learning, computing education, artificial intelligence This paper builds on a venerable tradition in the constructionist community, starting with “Twenty Things to Do With a Computer” that Seymour Papert and Cynthia Solomon published in 1971 as part of the MIT AI Lab Memos series. The memo presented a collection of different ideas and applications in the arts and sciences, and provided diverse entry points—some easy, others more difficult—into computing for children, youth, and teachers. The idea was to challenge conventions about what one could do with computers while also gaining deeper insights into one’s own thinking and computing. This tradition has been picked up when imagining things to do with programmable bricks (Resnick et al., 1996), with Scratch (Hideki, 2017), or with living materials in biology (Kafai & Walker, 2020). A recent book-length edition of essays (Stager, 2021) reviews these ideas fifty years later. Underlying all of these collections is a common thread of making things with computing rather than just observing or using them—an effort which we are continuing with a first attempt at twenty constructionist things to do with artificial intelligence and machine learning. In the last decade there has been an exponential growth in artificial intelligence and machine learning (AI/ML) applications, moving out of the lab into the world, impacting everyday lives. There is now a growing recognition that all students and teachers need to be prepared for understanding and using AI/ML applications (Long & Magerko, 2020; Touretzky et al., 2019) . However, to date most efforts promoting artificial intelligence in education have focused on what Eisenberg and colleagues (2017) called artificial-teachers, artificial-tutors, or artificial co-learners, centering learners as recipients of instruction from AI/ML agents or collaborators with AI/ML agents (Ouyang & Jiao, 2021) with much less attention paid to learners creating applications and designing culturally relevant and critically responsive AI/ML projects. Papert and Solomon (1971) observed that while most approaches to computing in education envisioned that “the transaction between a computer and the kid will be some kind of ‘conversation’ or ‘questions and answer’ in words or numbers,” (p.1) there are many other ways in which children could interact and create with computers. In Mindstorms: Children, Computers and Powerful Ideas (1980), Papert more clearly delineated two different visions for how children could interact and learn with computing, the instructionist one focusing on “the computer being used to p

FAQs

What does the paper build on, and what does it propose?

It builds on Papert and Solomon’s 1971 memo “Twenty Things to Do With a Computer” and proposes twenty constructionist things to do with AI and machine learning. Some ideas extend the original memo, while others are new.

How does the paper position constructionism in AI/ML learning?

It focuses on the constructionist vision in which children program and create with computing, gaining mastery and connecting to deep ideas from science, mathematics, and intellectual model building. This framing guides how learning AI/ML should be approached.

What kinds of AI/ML educational efforts does the paper critique?

It criticizes efforts that center learners mainly as recipients of instruction from AI agents (e.g., artificial teachers, tutors, or co-learners). It argues for giving more attention to learners creating applications and designing culturally relevant, critically responsive AI/ML projects.

Aug 5, 2026

Upload by Liam