Creative Interactions with Chatbots
An excerpt from "The Learner's Apprentice: AI and the Amplification of Human Creativity" by Ken Kahn
Maker educators are perhaps better positioned than most educators to understand the value of introducing AI to students to enhance their learning and actively develop new applications. After all, AI is a tool, a very powerful one, but conversations with AI chatbots can support student learning and increase their ability to do creative work.
Maker educators understand the importance of agency and that’s what AI has to offer.
“The advantage to A.I. that is, in some ways, a countervailing force here…is that it will increase the amount of agency for individual people.”
— Ben Buchanan to Ezra Klein on the New York Times podcast, The Government Knows AGI is Coming.
Recently, I interviewed Ken Kahn, author of “The Learner’s Apprentice: AI and the Amplification of Human Creativity” and the book’s publisher, Sylvia Martinez, for an episode of Make:cast, which I just published: “The Creative Potential for AI in Education.”
In the podcast, Sylvia Martinez remarks:
I think we've all seen how (AI) has exploded on social media and in articles and in books, but a lot of the uses are very utilitarian, how to make your grade book more efficient, how to write lesson plans, how to grade kids papers, how kids can use AI to write assignments, do assignments that are just old fashioned, work that they've always done.
And when Ken came to and both Gary (Stager) and me with this idea, it was so relevant for this practical aspect, what kids could do in the classroom, but also that the way he was doing it solved a lot of the problems that teachers were having with chatbots, the security issues, the sharing issues, having it write code that works in the browser is a tiny little twist that unlocks a lot of opportunity for use in schools.
Here is an excerpt from Chapter 1 of “The Learner’s Apprentice: AI and the Amplification of Human Creativity” by Ken Kahn, published by Constructing Modern Knowledge Press.
I believe with Dewey, Montessori and Piaget that children learn by doing and by thinking about what they do. And so the fundamental ingredients of educational innovation must be better things to do and better ways to think about oneself doing these things.
—Seymour Papert, from Teaching Children Thinking (1970)
BETTER THINGS TO DO
For over sixty years, these “better things” for learners to do have been computer programming projects. Seymour Papert, a former colleague of Jean Piaget and an MIT professor researching learning and AI, envisioned ways in which computing could change the nature of education. As early as the mid-1960s, he led efforts to introduce computer programming to children, resulting in Logo, the first computer programming language for children, and the precursor to the popular Scratch programming language used by millions of children worldwide.
A new way to do better things
The Learner’s Apprentice introduces an alternative way for learners to create “animations, simulations, and interactive games” (and much more) by conversing with chatbots like ChatGPT. While the day may come when artificial intelligence (AI) is capable of automatically creating whatever it is asked to make, for the foreseeable future learners need to “co-create” with AI.
This book presents examples of creating a wide range of very capable software applications (apps), illustrated stories, conversations with historical figures, text-based adventures, and much more. Using generative AI systems such as chatbots and image generators “can greatly expand the range of what and how children create” beyond what they can do today, even with programming tools such as Scratch or Python.
Today’s chatbots dramatically lower the barriers to creating games, stories, and sophisticated computer programs. But one may wonder—will the AI assist result in less learning? While researchers have yet to answer this definitively, there are strong arguments and anecdotal evidence suggesting that it can enhance the creative process. Students benefit most when they truly collaborate with the chatbot and not just ask it to do all the work.
Better ways to think about oneself doing these things
As Papert wrote in 1970, constructing things is only half of the story. Learning is best achieved by also reflecting upon the process of making things. How were problems detected and overcome? How were efforts split between planning, background research, tinkering with technology and ideas, building, sharing, testing, and fixing problems?
What worked well and what didn’t? Were dead ends encountered, and if so, how were plans revised?
In a school setting, assignments can encourage students to reflect as well as act. Chatbots can perhaps help if they are instructed to ask reflective questions at the right moments in a project. How to do this effectively and well is a research questions that is just beginning to be explored.
WHAT CAN YOU DO WITH A CHATBOT?
Chatbots can answer questions. They can engage in small talk. They help companies provide customer support, generate reports, and more. They can provide flash cards, quizzes, and the like to students. They can help teachers generate lesson plans and assess students.
But this book is not about any of these. It is about creative things you can do with a chatbot by your side. I’ve organized these things into three sections:
Make your own text-based adventure games, simulated dialogs, and virtual worlds.
Examples include witnessing the assassination of Julius Caesar, a foreign language learning game, conversing with historical figures, running a panel discussion, andparticipating in debates. (Section 2: Chapters 3 & 4)
Engage in storytelling and creative writing
Examples include stories about mathematical proofs and scientific phenomena, as well as stories about any imaginable topic such as “a puppy who goes to Jupiter on her birthday to solve a mystery and is a mermaid.” (Section 3: Chapters 5 & 6)
Create computer programs
Examples include making games and puzzles, creating tools like a customized calculator, programming ecological simulations or models of a solar system, creating augmented reality games, doing data science, and creating machine learning models. (Section 4: Chapters 7–16)
A novel approach to programming
Here is a software program that generates random nonsense words. I started with this prompt to ChatGPT:
Please make an HTML page that makes up new wordsChatGPT produced an HTML file that I could open with a browser. Every time I clicked on the “Generate Word” button it showed me a new word made up of three random syllables.
The code powering this interactive web app was generated by the chatbot without anything other than this one simple prompt. The conversation with the chatbot can then continue as new features are requested.
This, I believe, is a novel approach to programming. Using a chatbot as a partner in creating code offers a way to make interesting, interactive web apps without spending months learning to program. Some students may choose to explore programming in more depth, others might see making web apps as just a tool in their toolbox of useful and creative things they can do. Both are useful stances in this landscape of modern learning.
Throughout this book, the conversations involving the creation of computer programs serve these purposes:
Demonstrate the very wide range of apps that can be created.
Provide insights into how to guide chatbots to accomplish one’s goals. This includes providing feedback to the chatbot as the app is developed, asking questions,making suggestions, and helping the chatbot debug the generated programs.
Show how to deal with things when they go wrong.
WHAT CAN BE GAINED
As you co-create with AI, you learn how to incrementally develop things. The most effective way to create with an AI is to begin with a greatly simplified version of the desired end product. As you incrementally add more and more functionality, you hone your communication and design skills. Since chatbots make mistakes and misunderstand, you learn how to give effective feedback to the chatbot when things go wrong. While the chatbot may take over many of the low-level technical details, you still are the creative designer of the end product.
It is widely accepted that projects go better if students are passionate about them. Chatbots frequently respond very positively to a user’s suggestion before implementing it. Chatbots provide encouragement to persevere when things are going wrong. An open question is whether the typical positive and encouraging behavior of chatbots could play an important role in student engagement and learning.
There are many advocates for students learning to code. Yet there are not enough teachers trained in how to effectively teach programming, so many children miss outon this valuable experience. It is not possible in the foreseeable future that the goal of all children having experienced human tutors and teachers can be achieved. AI can change all this. Students can receive support from chatbots with their encyclopedic knowledge of many programming languages and tools. Age-appropriate explanations and support are available 24/7.
With today’s version of chatbots, anyone can create software applications by conversing in everyday language, without needing to learn the technical details of a programming language first. Apps co-created with a chatbot have a much wider range and more impressive capabilities compared to what non-expert programmers can typically create.
Chatbots have been trained with programs that use 3D graphics, process speech input and output, analyze data, incorporate pre-trained machine learning models for computer vision, natural language processing, network communication, and much more. This means they know how to create code to do all these things.
This is not to claim that AI is better than human teachers. There is no substitute for a caring, involved adult who is interested in what a young person can do, and who is equipped with knowledge and expertise to guide learners along the path to gain skills and knowledge. However, this can be a both/and conversation where students benefit from interacting with both teachers and chatbots.
The book is not one of those that debates the AI future. It is situated in the kinds of practical uses that you can do now. It is filled with lots of good examples of ways teachers and students can interact with AI in conversations.
In the podcast, I noted that “AI might be giving us another shot to change the learning experiences of kids.” Ken further illustrated this by explaining how AI can help students transcend the typical limits of school projects, pushing the boundaries to create more sophisticated simulations, games, and applications.
While there have been many attempts to do this in apps and websites by Edtech companies for twenty plus years, the results were often uninspiring. What happens when instead of being the user of apps or simulations, the students becomes an active participant in creating them and improving them?
For more information about the book, here’s is the publisher’s web page for “The Learner’s Apprentice: AI and the Amplification of Human Creativity