Final Reflection and Critical Evaluation
Project Statement
Project Title: AI and I: A Lazy Creator
Team Members: Bingrui Li, Xiaomeng Jiang
This project explored the possibilities and limitations of integrating AI into animation production. Through the use of tools such as ChatGPT, Seedance, and Kling, alongside traditional filmmaking methods, we investigated whether AI could function as a creator, collaborator, and critic within a contemporary animation workflow.
Rather than focusing solely on the final film, the project aimed to document the relationship between creators and AI systems, examining how AI influences visual development, storytelling, decision-making, and creative authorship.
My Contribution
Within the project, I was primarily responsible for environmental storytelling and scene development. This included storyboard design, prompt development, AI-generated environment shots, scene transitions, and the production of two major travel sequences.
I also conducted comparative testing between Seedance and Kling, exploring how different AI models interpret the same prompts. During post-production, I contributed to editing, colour grading, and visual integration to ensure consistency between scenes created by different team members.
Critical Evaluation
At the beginning of this project, we described AI as the “third member” of our team. However, after ten weeks of experimentation, I began to question whether AI can genuinely be considered a collaborator.
Although AI was capable of generating images, videos, and even providing feedback on creative work, it consistently failed to understand narrative intention. It could imitate cinematic language, but it could not explain why a particular shot should exist within a story. It could reproduce visual patterns, but it could not demonstrate genuine understanding of character motivation, emotional pacing, or thematic meaning.
This distinction became increasingly important throughout the project. The more AI-generated content we produced, the more human intervention was required to organise, evaluate, and contextualise the outputs. Ironically, the promise of automation often resulted in new forms of creative labour. Rather than reducing the role of the artist, AI shifted the artist’s responsibility from production to supervision, curation, and decision-making.
Another significant finding concerned authorship. If an AI generates an image based on my prompts, references, and creative direction, who is the author of the final result? Throughout this project, I came to view AI less as a creator and more as a probabilistic tool that operates within boundaries established by human intention. The creativity did not emerge from the AI itself, but from the process of selecting references, designing prompts, evaluating outputs, and constructing meaning from generated material.
Conclusion
This project ultimately challenged many of my initial assumptions about AI creativity.
Rather than proving that AI can replace artists, the project revealed how dependent AI remains on human judgement. AI can generate possibilities, but it cannot determine significance. It can create images, but it cannot understand why those images matter.
Therefore, I would argue that AI is not transforming animation by replacing creators, but by redefining what creative labour looks like. In an AI-assisted workflow, the artist becomes less of an image-maker and more of a director, editor, critic, and decision-maker.
For future research, I would be interested in exploring whether AI systems can move beyond visual generation towards a deeper understanding of narrative intention and emotional performance. Until then, meaningful storytelling remains fundamentally human.