In recent years, the world of art has undergone a significant transformation, with the incorporation of artificial intelligence (AI) into various creative processes. AI applications are now capable of generating poetry and designing visual masterpieces, breaking down the traditional boundaries of artistic expression. Whether you like it or not, this fascinating intersection of technology and creativity has brought about a profound impact on the art world.
Exploring the intersection of arts and science for real-world impact is at the core of ARISA, and today, we will discuss the research that we have been conducting to fulfil this mission, particularly in the field of dance and AI!
Robots can not only imitate human movements but also create captivating dance choreographies. This futuristic vision is now a reality, and AI has become a formidable force in the realm of dance, transcending mere imitation.
An interesting case in point is the Living Archives initiative, which employs AI to forecast and generate dance movements. This program uses a set of 25 years' worth of choreographies to predict new combinations of dance movements. It acts as an extension of an artist's imagination by creating novel and unique dance moves that go beyond what the human mind can conceive. This algorithm allows for unfathomable combinations of dance moves to be created, providing endless possibilities for creativity (McGregor, 2019).
Moreover, choreographers such as Wayne McGregor are utilizing AI as a tool for creativity, using it in all aspects of their work, including costume design and soundscapes. This mutually beneficial relationship between humans and AI is transforming the art of dance by opening up new avenues for artistic expression.
ARISA's researchers and other collaborators conducted two experiments to explore the relationship between AI and human creativity and perception. The experiments focused on understanding how different cues to human animacy change our perception and appreciation of dance. Human animacy refers to our ability to attribute human-like qualities to others, whether human or nonhuman. The ultimate goal of these experiments was to gain a deeper understanding of how AI and human creativity intersect in the domain of dance.
In the first experiment, researchers looked at how the dancer’s form – whether the dance was performed by a human avatar or a robot avatar – and the source of choreography – whether the choreography was human-generated or computer-generated – impacted how much the participants liked the dance video. The researchers also explored whether simply believing that a movement has been created using human motion capture (human origin) or using computer animation (computer origin) also impacts their aesthetic preference.
In the second experiment, researchers looked at how belief about the source of choreography impacted their aesthetic preferences. In reality, half of the participants saw only computer-generated choreographies but believed some of them were human-generated. The remaining half saw only human-generated choreographies but believed some of them were computer-generated.
These experiments revealed an intriguing pattern of results!
First, we humans really aren’t that great at knowing which choreographies are created by a human, and which are created by a computer! Curious to try it our yourself? Well, you can see the choreographies below and let us know which one you think is computer-generated, and which one is human-generated!
Were you able to make out the difference? It is probably easier for you if you are Bharatanatyam dancer! The participants who did our research were not trained in Bharatanatyam and therefore could not accurately tell which choreography was human- or computer-generated.
Yet, when they believed that a choreography was human-generated, they liked it more (even when it really wasn’t made by a human). And when they believed a choreography was computer-generated, they liked it less (even when it really wasn’t made by a computer). This shows that our beliefs play such a big role in our preferences – just our beliefs are enough for us to show an anti-AI bias, or a pro-human bias. This pro-human bias is often called anthropocentrism in psychology – a tendency to prioritize human concerns and perspectives above those of other organisms and entities! Says a lot about how we currently view AI. 😊
Yet, it wasn’t all negative for AI: our participants showed what is often known as agent congruence – that is, they liked a robot dancer when they thought the dance movement originated from computer animation, and when they believed the choreography was computer-generated. So, when the same category came together (here, all computer origins), people preferred that to an opposite category pairing (e.g., a robot avatar performing a human-generated choreography). Dance performed by robots was also likely to be watched more – perhaps because a robot dancing is novel, fascinating, and interesting for us, and we want to know more about it!
Finally, the researchers also found that attitudes toward AI (whether participants thought AI was useful, or whether they thought AI could lead to a dystopian future), and their experience with technology also impacted their preferences.
All these findings together provide valuable insights into the design and development of robotic dance by identifying how features of a dance choreography and audience demographics and characteristics influence the aesthetic experience of dance. Along with creating more refined creative productions by using AI in a dance maker’s toolkit, shaping people’s perception will be just as crucial!
The paper based on our research is currently in pre-print, and you can access it here! As you may know, ARISA Foundation has a podcast called Native Narratives, and the latest episode’s guest was Niranjan Pedanekar, a data scientist and a playwright, who discussed how he integrates AI in his creative process! If you thought this blog was interesting, you can check out our podcast on either YouTube or Spotify!
While it is an exciting leap towards unknown horizons, it is crucial to consider the ethical implications and societal impacts of this technological innovation. It is also essential to ensure that human creativity and artistic integrity remain at the forefront of our endeavours.
- Varun Kheria, Science Communicator, ARISA Foundation