This paper serves as a good starting point for us continuing research because it covers quite a broad range of topics.
The first thing that jumped out at me was that students were more likely to engage with the content when it felt more like a game than a school assignment. That fact will definitely shape the way I design whatever program I end up creating.
Another thing that interested me was the fact that students were so eager to learn about and discuss ethics and bias in AI. I had not even thought about the fact that many students either have not learned at all about bias in AI or have not been shown many concrete examples ("'when we typically think of Google, we think it’s objective, it’s always right'"). I will definitely look for ways to introduce and bring awareness to the issue of bias in whatever program I create.
The paper notes that a "sense of agency or control over technology has been long recognized as a key factor in impacting peoples' interactions with and perceptions of technology." I think that teaching kids about how exactly AI works and giving them hands-on experience with creating and controlling their own AI is a great way to give them this sense of agency and hopefully transform AI from an intimidating and nebulous new concept into an exciting area of new opportunities, which seems like it should be a main goal of whatever program I create.
And finally, this paper's broadness gives me the opportunity to look further into topics that it mentioned but did not go in-depth about, such as its use of a "participatory simulation approach."
This paper is really dense with information, and it is hard to tell where to start when discussing it. It obviously would serve as a great rubric when designing an AI Literacy tool (as we are doing), and it has a very large number of references for further research in many directions.
One thing that stood out was that the authors seem to have presented a potential project for us in the clearest terms: "Several recent projects such as Cognimates, eCraft2Learn, and others enable young learners to program AI. However, it is important for designers to keep in mind that prerequisite coding skills can be a barrier to entry, especially for children who are still learning how to read. Visual and auditory elements, fill-in-the-blank code, and Parsons problems are some techniques that can reduce this barrier. " It gives us examples of existing tools for children to learn to program AI, then states a major issue—that younger children without coding experience and with weaker reading and writing skills are unable to effectively use the tools—and suggests solutions to said issue. The idea is already laid out in this paragraph, and all that is left to do is create a new tool aimed at a younger and/or less experienced audience using the aforementioned techniques.
In all, this paper seems like it will be extremely useful as a sort of guidebook regardless of what kind of program I end up creating. It has a well-thought-out list of "competencies" required for AI Literacy which can be used as goals for teaching, and it also lists "design considerations" which can be used as guidelines for us to follow as we create our tools so that they are more effective.
For inspiration, we were instructed to utilize ChatGPT to "get some ideas for ways to engage children in learning about AI." I asked for help brainstorming, then I picked some of my favorites of the 50+ ideas I had it generate.
"Hello ChatGPT, I am in a course called "Developing AI tools for Children" and we have been tasked with brainstorming ideas for research-paper-worthy programs to teach middle-schoolers about AI. can you give me some ideas of what kind of program/tool to create?"
"Certainly! Creating educational programs to teach middle-schoolers about AI can be both exciting and beneficial. Here are some ideas for research-paper-worthy programs:
1. AI Storyteller:
- Create a program that generates interactive and personalized stories using AI algorithms.
- Introduce concepts like natural language processing and machine learning by allowing students to input their preferences and see how the AI adapts the story.
2. AI Ethics Game:
- Create a game where students navigate ethical dilemmas related to AI development and usage.
- The game could present scenarios where students make decisions about AI applications and witness the consequences.
- Emphasize the importance of ethical considerations in AI.
3. AI Virtual World Builder:
- Create a virtual world-building tool where students use AI to design landscapes, characters, and stories.
- Introduce concepts like procedural generation and AI-driven storytelling.
4. AI in Gaming:
- Create a program that teaches students how AI is used in gaming.
- Explain concepts like pathfinding, decision trees, and reinforcement learning by allowing students to modify and improve the behavior of in-game characters.
5. AI in Everyday Life:
- Develop a program that showcases AI applications in daily life.
- Highlight how recommendation systems, voice assistants, and image recognition are used, making AI concepts more relatable to students.
6. AI Detective Game:
- Create a detective-themed game where students use AI tools to solve mysteries.
- Introduce concepts like pattern recognition, data analysis, and problem-solving in the context of a detective story.
7. AI in Healthcare Simulation:
- Create a simulation program that demonstrates how AI is used in healthcare for diagnosis and treatment.
- Allow students to virtually experience medical scenarios where AI plays a crucial role.
8. AI Coding Challenges:
- Develop a platform with coding challenges related to AI concepts.
- Allow students to solve problems and see the impact of their code, reinforcing programming skills alongside AI understanding.
These ideas blend educational content with interactive and engaging experiences, making AI concepts accessible and enjoyable for middle-schoolers."