Co-Creation with Generative AI in Video Production
YUSHIEN, 2025
Abstract
Generative AI has rapidly emerged as a transformative force in
creative industries, enabling new forms of collaboration between
humans and machines. Recent advances in AI models for text,
image, and video generation have sparked widespread interest in
marketing and advertising. Brands and creators are experimenting with AI to streamline content creation and inject fresh creativity into campaigns and video productions.
At the same time, creating ad content entirely with AI is a bold endeavor fraught with uncertainties. Not all generative AI marketing efforts have been well-received. For example, in 2023 Toys“R”Us released what it touted as the first brand film made with OpenAI’s text-to-video tool Sora, but the experiment received public backlash (Placido, 2024). Likewise, Coca-Cola’s 2024 holiday ad, which animated with AI, sparked debate and mixed reactions from consumers, with some viewers deriding it as lacking
the human touch ( Katie Deighton and Megan Graham, 2024).
These incidents underscore both the promise and the perils of AI-driven creativity. In this context, I undertook a project to investigate generative AI co-creation in a commercial campaign setting. The project’s aim was to create a promotional video for OCAD University that is “bold, slightly bizarre, and wildly imaginative,” leveraging a suite of AI tools across modalities (text, image, audio, and video). This endeavor serves as case study of human–AI collaboration in the end-to-end production of an advertisement. The central questions explored include: How can various generative AI tools be orchestrated in a creative workflow? What advantages and limitations aris from using AI as co-creator, compared to traditional production? How does the resulting creative output reflect the influence of AI on style and content? And In a collaborative context with AI, how can we prevent the work from losing its human touch?
marketing and advertising. Brands and creators are experimenting with AI to streamline content creation and inject fresh creativity into campaigns and video productions.
At the same time, creating ad content entirely with AI is a bold endeavor fraught with uncertainties. Not all generative AI marketing efforts have been well-received. For example, in 2023 Toys“R”Us released what it touted as the first brand film made with OpenAI’s text-to-video tool Sora, but the experiment received public backlash (Placido, 2024). Likewise, Coca-Cola’s 2024 holiday ad, which animated with AI, sparked debate and mixed reactions from consumers, with some viewers deriding it as lacking
the human touch ( Katie Deighton and Megan Graham, 2024).
These incidents underscore both the promise and the perils of AI-driven creativity. In this context, I undertook a project to investigate generative AI co-creation in a commercial campaign setting. The project’s aim was to create a promotional video for OCAD University that is “bold, slightly bizarre, and wildly imaginative,” leveraging a suite of AI tools across modalities (text, image, audio, and video). This endeavor serves as case study of human–AI collaboration in the end-to-end production of an advertisement. The central questions explored include: How can various generative AI tools be orchestrated in a creative workflow? What advantages and limitations aris from using AI as co-creator, compared to traditional production? How does the resulting creative output reflect the influence of AI on style and content? And In a collaborative context with AI, how can we prevent the work from losing its human touch?
Keywords: Generative AI; Human-AI Co-Creation; Multimodal AI; Text-to-
Video; Video Production.
Click to read the paper