I found these readings extremely enlightening about the holistic use of AI in museums. In general, I believe that AI and ML can be used as a buzzword in today’s technology sphere and inventions, it comes to indicate this all-knowing purpose of the future and all-doing system. As the Styx article mentions, many people tend to think of AI as these robot helpers that can interact with visitors, a replacement for a human. I personally have also tended to judge AI with a negative bias because of the promises that tend to come with the buzzword. However, these articles bring up useful and advanced examples of applying AI to solve problems within the backend of museums.
However, AI is still a highly tailored system that needs to work on specific datasets and in controlled environments, which museums provide through their collections. Since these collections are comprehensive and internalized, and often also come with embedded related knowledge already, this is the perfect place for AI to thrive. One massive generative AI piece that I have thought about often is the Refik Andol: Unsupervised piece in the Moma. It is a large-scale constantly generating show based on the museums’ collections. It is a fantastic piece to watch in action, and also has an interesting display that explains to visitors how AI is made. I also believe generative AI can be used in the art form to be a strongly spoken criticism to how AI is treated as an objective creation, when it is truly ingrained with many biases. This brings into question whether it is good to have truly unsupervised generation, or if there needs to constantly be human oversight. I agree with the CHM paper that the project needs to be well defined and with careful requirements.
My final thought is that AI, currently a very text and 2-d image centric, especially even 2-d generation is only picking up in a widespread manner now with Dall-e and other image generator prompts. I am excited to see what 3-d AI generation could bring, not just computer models, but how it has and can change fabrication techniques, and consequently human-computer interactions, in the future.
(Reposted ?? I’m having trouble publishing my readings on github they don’t seem to be consistantly going through)