The world has been shaken up by generative Artificial Intelligence systems that can engage in conversations on any topic and generate images, songs, and video from text requests. 400 million people use ChatGPT, the most popular chatbot, at least once a week. But what about school? Chatbots are being used regularly by many students. Too often they are asking the chatbot to do their work for them, learning very little in the process. Is AI in education destined to deprive students of the experience of creating, researching, and thinking things out themselves?

Some schools are thoughtfully exploring how AI can enhance student learning while others are reacting with outright bans or purchasing tools meant to “catch” students using AI. But many schools are doing little to help students and teachers navigate the various questions that AI use brings up.

Meanwhile, books and articles about AI in education tend to mostly cover uses of AI as lesson plan generators, worksheet makers, and other administrative tools for teachers. For students, AI may be recommended as simply a glorified search engine with warnings that students should be prevented from cheating on assignments. The recommendations tend to reinforce the worst aspects of school, from five paragraph essays that no one will read to worksheets, quizzes, and rote memorization.

But instead, imagine that chatbots can be used to augment human creativity, curiosity, and problem solving.

This can be done with today’s chatbots, if they are treated as partners in creative projects. Students can focus on design, critical thinking, communication skills, and problem solving. They can leave the low level technical details to the chatbots if so desired. In doing so they can embark on much more ambitious projects than if they worked solo.

My journey exploring how children can creatively use computers, and especially AI systems, started when I was a graduate student at the MIT Artificial Intelligence Lab fifty years ago. My doctoral thesis involved an AI system that generated computer animations from story descriptions, anticipating some aspects of today's text-to-video AI models. In the time since, I’ve been an advocate for creative computing as a computer scientist, researcher at Xerox PARC, software developer, and professor of learning and computing or researcher at universities from Sweden to Oxford to Singapore.

Any chatbot can be a partner in creative projects, helping students “learn by doing” about how AI can help with intellectual work in the modern world.

Make software – without learning to code

Any of today’s chatbots can create code for a wide variety of web apps. Since ChatGPT debuted, I’ve developed hundreds of examples of how students can express their creativity by co-creating apps with chatbots. They range from simple apps a young child can make to high school level projects creating complex scientific and machine learning apps. There are ways to create effective prompts to accomplish learning objectives using guiding prompts and custom GPTs. But even with simple prompts, teaching students to guide a chatbot through the process of creating an app, and then making it better, is an exercise in critical thinking and iterative design.

I advocate using chatbots to co-create “web apps” as a particularly effective way for students to create interactive browser-based applications that are easy to share and run safely on any school or home computer. Browsers are already configured for student use in most schools, so building apps to run in the browser addresses a myriad of security and accessibility concerns without additional software installation.

Examples of apps that students can co-create with chatbots include:

• A wide variety of games, from word games to graphic adventures
• Scientific explorations and simulations
• Augmented reality experiences
• Incorporating AI into apps
• Mathematical explorations from proofs to data science
• Using and training your own machine learning models
• Programming microcontrollers and mobile devices

Create history, science, and literature adventures

Beyond app development, generative AI opens up new possibilities for creative exploration across all subjects. Students can use chatbots to create historical text-based adventures, simulate debates and panel discussions, and create illustrated stories. Due to the encyclopedic knowledge of chatbots these can be on any subject and in any depth. Teachers can learn how to customize chatbots to have the desired pedagogy, tone, and complexity of language via carefully crafted prompts.

Examples of creative conversations and adventures include:

• Participating in a conversation with Aristotle, Galileo, and Newton on gravity
• Conversing with a lion and an elephant about their lifestyles
• Moderating a simulated panel on generative AI and education
• Joining philosophical discussions while visiting ancient Athens
• Commanding a spaceship in French in an unfolding story about a mysterious artifact

Enhance creativity and storytelling

Write and illustrate stories across different styles, time periods, and subjects, and connect creative writing to various disciplines. While you can ask a chatbot to do creative writing, the result and the learning will be much greater if you treat the chatbot as a collaborator. You might begin by brainstorming ideas with the chatbot followed by co-creation of the outline. When writing fiction, you might bounce around ideas for the plot, characters, and dialog. The chatbot can critique and make suggestions to your drafts or you can give comments to its drafts.

Examples of stories co-written with AI describe:

• How arranging pebbles leads a prehistoric girl to discover Euclid’s proof that there is no largest prime number
• A bird’s first person story about natural selection
• A heroic narwhal who is afraid of heights who must climb Mt. Everest to retrieve a magical amulet
• Accommodating infinite numbers of guests to Hilbert’s Infinite Hotel
• A time traveling high school student who visits Mesopotamia as writing is being invented

Any imaginable literary form can be co-created with a chatbot: fiction, poetry, songs, scripts, and much more. These can be connected to the curriculum in any subject.

Supports project-based learning

Underlying all these activities is the pedagogical educational theory of Constructionism pioneered by Seymour Papert. Constructionism emphasizes the learning that occurs when students do self-directed projects. Learners are encouraged to tinker with technologies such as chatbots. An important part of constructionism is acquiring reflective skills while doing projects. It encourages sharing products and the process of their creation.

Debugging – a critical thinking skill

Chatbots make mistakes. They often hallucinate facts. Students learning to be critical and fact checking chatbot responses is a valuable skill. Chatbots also sometimes produce buggy code. Students become digital detectives, fact-checking and debugging alongside their AI partners.

Isn’t it cheating?

Students who ask AI to do all their work for them are cheating themselves and missing out on important learning opportunities. In contrast, students who treat AI as a partner can, not only accomplish much more, but have many opportunities to improve their communication, critical thinking, and design skills. Students will not “cheat” if they are assessed not only on what they produce but also on their reflections of their process of co-creating with AI.

Collaboration with chatbots expands the range of what students can do

AI collaboration expands what's possible - students can tackle more ambitious projects while focusing on creativity and problem-solving rather than technical details. They can create apps that incorporate advanced capabilities such as speech, vision, machine learning, and 3D graphics without first mastering the otherwise necessary technical details. They can co-author sophisticated multimedia projects taking on some of the creative roles and leaving the remainder to generative AI. In partnership with AI students can create projects that solve real-world problems or contribute to science and creative arts.

About the author

Ken Kahn, PhD is the author of a new book, The Learner’s Apprentice: AI and the Amplification of Human Creativity. Kahn brings a wealth of experience and expertise to this revolutionary work. After getting his doctorate at the MIT Artificial Intelligence Laboratory, Ken has been an advocate for creative computing as a computer scientist, researcher at Xerox PARC, software developer, and professor of learning and computing. The book’s examples document a thoughtful exploration of the untapped potential of AI chatbots and are shared with lively anecdotes and practical lessons for readers to try themselves. Read about ongoing research, or follow Ken on LinkedIn or Facebook by visiting tinyurl.com/ken-kahn-home-page.