This paper introduces ChAI, a collection of interactive machine learning tools for the ChucK music programming language, and chronicles its use in Music and AI, a critical-making course at Stanford University. We explore the intersection of AI and music HCI, emphasizing the augmentation of human creativity, rather than its replacement. We introduce ChAI's interactive AI tools and toys and highlight philosophical, ethical, and cultural considerations. Focusing on interactive machine learning, audio analysis, feature extraction, and audio re-synthesis, this paper shows how ChAI could be used to create playful, expressive systems (in an age of generative AI that tends to limit human interaction in favor of prompt-based automation). In the critical-making course, ChAI was used alongside reflections on AI's broad impact on the arts and who makes art, while stressing the importance of understanding AI's societal and cultural implications. This work not only introduces a new tool but also invites broader conversation on the role of HCI, play, and "small-data" approaches, in music and AI practice and education.