TLDR: A study investigated how informing audiences about the use of generative AI in a live dance performance affects their perception of its artistic value. It found that audiences were more inclined to attribute artistic merit to AI-generated elements when they were unaware of AI’s involvement until after experiencing the performance, suggesting that transparency about AI’s presence, rather than its mechanics, is crucial for artistic reception.
The integration of artificial intelligence into creative fields, particularly art, has sparked considerable debate regarding the value and perception of such works. A recent study, titled “What’s Behind the Magic? Audiences Seek Artistic Value in Generative AI’s Contributions to a Live Dance Performance”, delves into how audience awareness of AI’s involvement influences their appreciation of art.
Conducted by Jacqueline Elise Bruen and Myounghoon Jeon, this research aimed to understand audience perception under two conditions: when information about the technology used was withheld, and when different types of technology (Generative AI versus traditional digital tools) were employed.
The Experiment Setup
The researchers developed two distinct versions of a live dance performance, each augmented by technology. One version incorporated Generative AI (GenAI) tools, while the other utilized traditional digital tools. To explore the impact of information, audiences were divided into groups: some were informed about the technology before the performance, and others were told only after completing a survey about their perceptions.
Thirty-nine participants were involved, distributed across four performance scenarios: AI/Told Before, AI/Told After, Non-AI/Told Before, and Non-AI/Told After. The performances themselves were identical in choreography and structure, lasting 12 minutes and divided into three scenes. Live data from the dancer’s breathing and location, captured by sensors, dynamically altered visuals on a large screen and created sonification (sound representation of data).
In the Non-AI performances, a technologist made all creative decisions regarding imagery and data mapping, sourcing visuals from human-produced works. For the AI performances, Generative AI (specifically ChatGPT) was used to design imagery and data mappings. Additionally, neural network models, built with Fiebrink and Cook’s Wekinator, were employed in one scene to interpret dance subroutines and influence the visuals and sound.
Key Findings on Audience Perception
The study revealed fascinating insights into how audiences perceive technologically augmented art:
- For performances using traditional digital tools (Non-AI), audiences who were informed about the technology beforehand were more inclined to actively look for patterns between the visuals and the dancer’s movements. They also found the visuals more distracting compared to those who were told after the performance. This suggests that prior knowledge can heighten attentional focus on the technological elements.
- In contrast, for performances incorporating Generative AI, audiences who were informed about AI’s use only after the survey rated the work significantly higher on artistic merit. Conversely, those told about AI before the performance were more likely to perceive the projected visuals as random. This indicates that an audience’s awareness of AI’s involvement can significantly influence their judgment of a work’s artistic value, often leading to a more critical view if known upfront.
- When comparing audiences who were told about the technology after the performance, those who experienced the AI version expressed greater curiosity about the artistic decisions made. Meanwhile, those who saw the Non-AI performance rated the complementary nature of the sound higher, suggesting they could more easily draw relational conclusions between elements in the non-AI work.
Also Read:
- Bridging the Cognitive Divide: Why AI’s Goals Differ from Human Intentions
- Unlocking the AI Black Box: A New Framework for Transparent and Personalized Learning
Implications for Artists and AI Explanation
The research highlights a crucial point: the way a creative work is presented and the audience’s knowledge of its production method profoundly impact their perception of its artistic value. The findings suggest that for GenAI art, transparency about its presence and power might be more beneficial than detailed explanations of its technical mechanics, especially given that many GenAI tools operate as “black boxes.”
This study encourages artists to consider how they present their work, particularly when incorporating AI. It also calls upon the Explainable AI (XAI) community to shift focus towards explaining the *presence* and *impact* of GenAI in artistic contexts, rather than solely on its internal workings. Understanding these audience biases can help bridge gaps in understanding and foster a more nuanced appreciation of AI’s contributions to art.
