Why Project Based AI Learning? A Guide for Parents

This is the fourth article in our “Guide for Parents” series about why and how K-12 students should learn AI. Previous articles covered the reasons to learn AI, the relationship between AI and Robotics, and the relationship between AI and Programming. In this article we focus on how to learn AI, and in particular why project based learning is so helpful for K-12 students to both learn and enjoy AI.

Our Experience with Projects

At http://aiclub.world, we have seen students from Grades 5–12 build hundreds of AI projects to solve everything from predicting how much life is left in your phone battery to detecting disease in plants. They have gathered data from sources as varied as the Centers for Disease Control, the World Health Organization, the stock market, population census, public data repositories, community surveys, movie reviews, the internet, and their own experiments. You can see some of their projects here and see a live project presentation here.

What we have learned from this experience:

  • Projects are a great way to not just learn AI but develop a lifelong passion for solving problems with technology. Many of our students have gone forward to build additional AI projects, use AI for STEM projects in their schools, use AI for community projects, and to compete and win in AI competitions.
  • Students bring amazing creativity to projects! Most of our student projects are their own ideas, where they have come up with new and innovative ways to apply AI.
  • As students define their project, they learn a lot about how to apply AI to problems, when to use AI, and when not to. As they gather data for their project, they learn how to respect privacy and data ownership, how to distinguish between good data and bad, and how to apply AI ethically and fairly.

Beyond these observations, there are three reasons why we focus on project based learning:

Projects Make AI Learning Accessible and Fun

AI has been traditionally hard to learn, even for adults. There were several reasons for this. First, the tools available required deep knowledge of programming, and sometimes math, to use. Second, building an “end to end” AI, where you can easily ask it questions and get answers, was even harder because it required the user to connect several tools. Third, most AIs had to be tuned by hand, making it even harder for a beginner to get something working well without learning a lot about algorithms.

Fortunately, now, with tools like http://aiclub.world, these are no longer issues:

  • Anyone can build their first AI in 10 minutes or less, without doing any programming or applying any math. In 10 minutes, they can have a working AI that they can directly ask any question they like and get an answer.
  • The AIs can then be tuned as much or as little as the user wants. The user can then learn progressively, one step at a time, and see their AI improving as they apply more advanced concepts. They can also build fancier AIs and applications as their programming and math knowledge increases.

This opens up an entirely new way for K12 to get excited about AI. Students can build an AI project to answer questions they care about, and see their AI learn and improve right in front of them. In our experience, students spend far more time improving their AIs than we ask, each time being driven by their curiosity and ownership pride in their creation. They are motivated to learn more about how the AIs work and apply those insights to improve their project.

Project Based Learning teaches Practical Concepts that are Core to Successful AI

In the professional world, most AIs follow a lifecycle (see example below) going from problem to solution to improvement. To get an AI working well in real life requires the human to combine several things: improving data, improving the AI learning, and iterating based on feedback from real world use. Professionals who have this experience of seeing the AI actually applied to solve a problem (frequently called Full Stack Data Scientists) are far more in demand than those who know just parts of the solution.

The Lifecycle of a Typical Artificial Intelligence Solution

By doing projects from day one, students are practicing the same fundamental principles. They learn critical concepts like the need to manage data well, understand the source of the data, appreciate the relationship between data and algorithms, and so on. Project based learning connects AI concepts and theory to solutions. Students do not just learn concepts in a vacuum. With every concept they learn, they can immediately apply it directly to their problem.

Projects help a Student’s Future, in AI and Beyond

AI is a rich field with many many powerful techniques. One of the most valuable things that AI professionals do is find the right AI techniques to solve specific problems, and decide how and when to apply programming or other STEM skills to improve the solution. These experts can also recognize problems that can be solved by AI, and distinguish them from problems that should be solved in other ways. They understand that AI is not magic but rather a tool that sometimes works well and sometimes does not. These are skills learned from experience and from using AI to solve lots of problems. Project based learning helps students appreciate these nuances and master the tools of AI in the same way.

Projects are also fundamentally solutions to problems. Students learn how to think about a problem, how to assess the quality of a solution, how to improve the quality, and then describe their solution to others with a focus on the decisions they made and why they made them.

How to Get Started?

How can you get started with project based AI learning? You can sign up for a free account at http://aiclub.world and explore projects on your own, or sign up for one of our K12 courses https://aiclub.world/workshops.

I am a software engineer by training and an expert in artificial intelligence. I am also a mom! I teach coding and tech to kids from grades 4-12.