Experimental composer Holly Herndon built an AI voice clone that anyone can use

Scientific American
Holly Herndon creates AI models of her voice, believing AI enhances collective creativity and offers new artistic possibilities.

Summary

Experimental composer Holly Herndon is pioneering the use of AI to expand creative possibilities, developing custom models that allow anyone to perform using her voice. Her approach, termed “protocol art,” focuses on creating the rule sets for art generation rather than the final media itself, emphasizing the process of dataset creation and model training as integral artistic acts. Unlike commercial AI models trained on broad internet data, Herndon builds her own datasets, resulting in unique and less constrained outputs. A key project, Holly+, allows real-time voice cloning, enabling others to sing with her voice, even adopting different accents. Herndon envisions a future where AI facilitates real-world collaboration and collective intelligence, moving beyond the limitations of individual creativity and the attention economy, and ultimately redefining our understanding of creativity itself.

(Source:Scientific American)