||An introductory lesson to basic visual AI and computer vision. It goes over the concepts of movement and how an AI application will learn to identify it, using MIT Media Lab PoseBlocks/Scratch as well as a live demo of Google's Quick Draw. It also contains short activities to demonstrate weaknesses in modern AI and AI ethics challenges. Includes lesson plans for middle school and high school levels.
||Computer vision, machine learning models, visual AI, adversarial data, AI bias.
||K-12 (is utilized as an introduction to AI)
||Low difficulty throughout major duration of the lesson - during free-project time, the difficulty may increase slightly depending on the student. Intended to be completed in three days.
||Allows for a live demo of working AI, which means that students experience a theoretical and practical introduction to ML models and features before getting to create their own. The high school level allows students to create a fast and easy custom ML model and import it into a Scratch game environment, while the middle school level allows students to jump into Scratch AI programming with pre-created models. This allows for a range of difficulty for students via a well-known block-programming platform that emphasizes ease of use.
||Only provides a very basic introduction to AI concepts, and doesn't leave much room for personal growth with knowledge following the lesson. Students can continue to create their own models and Scratch applications based on them, but the threshold of what they can further experiment with beyond the lessons is somewhat low. The lesson does not easily bridge students to prepare for more complex models and work in higher AI subjects.
||All aspects of the lesson are browser-based, and can be completed with access to a sufficient Internet connection on any operating system. Block-programming is the only form of programming langauge utilized. Ideally, there is a prerequisite knowledge of Scratch before beginning the lesson.
||Instructors can vary the assignments by introducing alternative AI projects to lead students through, or alternative Teachable Machine models to create online. They can also center the lesson around a variety of themes that suit their interests, changing the resulting models and Scratch applications taught to match those themes. In particular, for social impact, an instructor can focus on such models with a potential impact (ie. models that can identify recyclables from non-recyclabes, etc.) and focus the lesson on themes such as creating impactful applications through AI models.