Experimental Animation in the age of AI

With the rapid advancements in image synthesis by Artificial Intelligence, animation as a digital art form is encountering unprecedented changes in its production pipeline, aesthetics, accessibility, and authorship. Machine learning algorithms can analyze large datasets of existing animators' styles to identify patterns and then use this data to create new animations in similar or modified styles. Furthermore, AI has shown its capability to suggest complementary movement, composition, color, and sound as a new form of creative collaborator. One of the most up-and-coming AI imaging technologies is synthesizing videos guided by text or spoken dialogues in real-time using natural language processing. In these new movements, animation seems to no longer require time-consuming tasks frame by frame. Despite these advancements, the temporal consistency and quality control issue of AI-generated animation are still a concern to conquer. It is because the current AI imaging studies focus mainly on integrating AI into industry animation projects. However, there are many possible creative ways of unleashing AI tools for experimental animation projects. In this presentation, the artist will showcase their new AI animation projects, Dissolution (2023) and Cunabula (2023), to demonstrate the possible experiments with generative adversarial networks and neural image stylization techniques to create abstract and surreal animation sequences. The presentation will provide insights into the possibilities and limitations of AI in independent animation practice, emphasizing how artists can leverage this technology to enhance their creative process while retaining their position as primary investigators in the new forms of collaboration in the age of AI.