The Salzburg Festival on Aug. 15 played host to a discussion on predictions, opportunities, and doubts about the intersection of artificial intelligence and art, including a performance by star soprano Asmik Grigorian and an AI model of her voice.
The discussion, AI & Art, was organized by the CultTech Association, a Vienna-based non-profit focused on educating and shaping the startup ecosystem around new technologies that benefit culture and society. In addition to Grigorian, panelists included theater director Stefan Kaegi, music professor Miller Puckette, entrepreneur David Yang, artist Gerfried Stocker, and consultancy founder AC Coppens. An AI called “Morfeus” also joined the conversation.
Grigorian was the event’s highlight, performing the medieval Armenian song “Krunk.” The same song was also performed by an AI trained on Grigorian’s voice with surprisingly realistic results. The commercial tool Voice-Swap.ai was used to create the model.
“When I heard my AI-generated voice, my first reaction was to laugh,” said Grigorian. “I expected to hear more of myself in it, and what I heard was a very beautiful, great performance, but it has nothing to do with the temperament of my voice, I could hear nothing of myself. I understand that I started to see the artificial intelligence in art more as a piece of art itself, because it is created by human beings with their own fears, with their own dreams, with their happiness and pain.”
The recent intersection of AI and art is not without controversy. The Recording Industry Association of America (RIAA) has sued AI song generators Suno and Udio for copyright infringement, while a group of authors led by comedian Sarah Silverman have sued AI leader OpenAI for copyright infringement, with only limited success so far.
