Reading Response #9

From Ge Wang's "Humans in the Loop: The Design of Interactive AI Systems" and Allison Parrish's "Experimental Creative Writing with the Vectorized Word," I would like to respond to one of the shared central ideas: whether artificial intelligence makes people more creative or less creative. Ge's essay and Allison's presentation give two opposite examples. In Ge's essay, he mentions the technology of transferring artistic style and the loss of meaning while infusing the styles of thought-provoking pieces onto photos with completely unrelated themes. On the contrary, Allison applies the approach of vectorizing words with respect to their meanings to the phonetic space and combines it with the processing methods previously used on audio and visual, resulting in an unprecedented genre of poems.

The same debate occurs in the music field as well. Based on determined rules, it is feasible to list all possible melody and chord progression with the help of computer, but does that mean instead of creating, composers are just selecting pieces? Recently, Google launched an application called Tone Transfer which is able to transfer a certain melody from an instrument, including human voices, to any other instrument. This sounds similar to artistic style transfer but leads to apparently more positive comments. As western music and instruments dominate the music trend, there still exist a lot of domestic music customs that do not fit into western scales. Tone Transfer enables musicians from all over the world to express their traditional music with a more widely accepted medium and may inspire them to produce new or fused types of music.

I really like the statement from one of my HAI professors: "Music is not just influenced by technology. Music is also heavily influenced by culture." I believe this also applies to the visual art and poetry domain. Technology is just a tool which opens more possibilities for creative people, and itself cannot kill creativity. Though it is the responsibility of the designers to design tools that mainly focus on approachability and experimentation, different ways of using these tools induce varied products as well. As for neural style transfer, it can actually be used to understand how human beings appreciate art. To illustrate, if we extract some intermediate layers from the neural networks, we can see how a particular layer is trying to identify specific objects like dogs. Further investigation may contribute to the understanding of art and help evoke even more creative pieces. Let technology, especially AI, be the catalyst of creativity.