Although constructionism originated embedded within early artificial intelligence research, it is only recently that researchers have returned to designing and researching constructionist tools and activities for learning Artificial Intelligence/Machine Learning (or: AI/ML). The pervasiveness of AI/ML in the everyday lives of young people—impacting how they connect with friends, listen to music, play games, or attend school—coupled with the accessibility of large language learning model applications, discussions about algorithmic justice, and growing efforts to incorporate computing into K-12 education increase the urgency of AI/ML education. Yet, within education, most AI/ML efforts have centered on learning analytics and providing scaffolds to learners, focusing on what Papert called “the computer being used to program the child.” In contrast, constructionist AI/ML efforts center on designing learning environments and researching how young people can create personally relevant AI/ML powered applications. In this symposium, we bring together historical perspectives on constructionism and AI/ML, examine current efforts that build on learner’s interests, and develop possible directions for future research and design. We discuss how when creating AI/ML powered projects, teachers and learners can collaboratively develop and integrate conceptual and critical understandings that are increasingly important to participate in the world.