Workshop on Interactive AI Systems for Live Audiovisual Performance

The workshop provides a valuable perspective on the intersection of AI and music, challenging the dominant narrative of generative AI as a shortcut to creative output. Ge Wang’s critique of Suno’s CEO statement is well-articulated, drawing an analogy between music-making and hiking to argue that the process, with its challenges, is integral to the meaning of the activity. This argument is compelling and grounded in a humanistic understanding of creativity. Celeste Betancur’s live demonstration effectively illustrates the concept of human-in-the-loop AI, using a minimal HMM model to generate rhythmic patterns in real-time. This approach highlights the potential for AI to serve as a collaborative tool rather than a replacement for human creativity. The technical details are accurate, and the references to tools like ChucK and Wekinator are credible. However, the discussion is largely opinion-based and lacks empirical evidence or systematic analysis. The workshop is more of a philosophical and practical exploration than a rigorous scientific study. The sources cited are primarily the speakers’ own work and well-known tools, but no external references are provided. The title accurately reflects the content, and the workshop is well-structured, with a clear demonstration and discussion. Overall, the content is thought-provoking and relevant, but its scientific rigor is limited by its anecdotal nature. The absence of a formal evaluation of the tools or methods presented is a notable weakness. Nevertheless, the workshop offers valuable insights for practitioners and researchers interested in creative AI applications.

The title accurately reflects the content: a workshop on interactive AI systems for live audiovisual performance, with demonstrations and discussion.

8/10

The workshop features expert practitioners (Ge Wang, Celeste Betancur) from Stanford CCRMA, demonstrating live coding and small-scale ML models. The discussion is grounded in practical experience and references to established tools (ChucK, Wekinator) and researchers (Rebecca Fiebrink). The content is technically accurate but primarily opinion and demonstration, not peer-reviewed research.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 00:00Introduction and HAI Preamble read by Ge Wang
• 05:00Discussion on AI and music, comparison to electric guitar and email
• 10:00Ge Wang reads from his article 'GenAI Art Is the Least Imaginative Use of AI Imaginable'
• 15:00Introduction of guest Celeste Betancur and her live demo
• 20:00Celeste explains her use of ChucK and HMM model for live coding
• 30:00Discussion on human-in-the-loop AI and small data sets
• 40:00Q&A session with audience
• 1:00:00Further demonstrations and discussion on tools like Wekinator

• ChucK programming language — Mentioned as the primary tool used for live coding and audio programming.
• Wekinator — Mentioned as a tool for interactive machine learning.
• Rebecca Fiebrink's work — Cited as inspiration for the human-in-the-loop approach.

• ChucK — The tool is widely used in computer music and aligns with the workshop's approach.
• Wekinator — Supports the idea of accessible machine learning for artists.

• Suno — Ge Wang criticizes Suno's approach to AI music generation, contrasting it with the human-in-the-loop philosophy.