GPT
-
January 9: AI and Performance
Natural Performances with Artificial Intelligence
Using GPT for Scenes, Imagery, Focal Points and Emotion
Paper by Jonathan Robie and Discussion on Zoom
Thursday January 9, 2025
2pm Central Time
Join us for presentation and discussion of Jonathan Robie's "Natural Performances with Artificial Intelligence: Using GPT for Scenes, Imagery, Focal Points and Emotion" on Zoom on Thursday January 9, 2025 at 2pm Central Time. A performer embodies the text, using voice and gesture to convey emotional flow, imagery, relationships in time and space, and other aspects of the text. Accuracy requires detailed, careful analysis of the text. The performer must study, imagine, and internalize these relationships. This presentation uses GPT (systems like ChatGPT) to prepare a Psalm for performance by: 1. Dividing the Psalm into distinct scenes. 2. Highlighting key imagery within each scene. 3. Mapping the emotional progression throughout the Psalm. 4. Creating questions to aid deeper understanding and internalization. We use a GPT framework to analyze biblical texts in Hebrew or Greek. A user can specify an analysis and the desired output structure, applying it to a set of passages. It is easy to try out new analyses, weeding out unhelpful ones and applying useful analyses to other texts. The analyses mentioned above can be found here for each of the Psalms: https://docs.google.com/document/d/1JHHwtbKeFeOmHl0-wxOFqE_5W9DgmpbWiYWtn0QBqFM/edit?usp=sharing Emotional Flow uses a model of emotion closely related to "heart" in both Greek and Hebrew, based loosely on Martha Nussbaum's "Upheavals of Thought." Emotion enriches our relationships with God and others, guiding us in situations beyond our control and understanding, and helping flourish. These GPT-based analyses surprised us. They reveal deeply human aspects of biblical texts, offering multiple perspectives to help performers internalize and embody the text. We end the presentation by doing an oral performance of the Psalm based on these analyses.
Register for the Zoom meeting at https://us02web.zoom.us/meeting/register/tZwtf-GrrDgqE9EIv89WtgqPrgdk9AtsEcA_