Why This Topic?
Virtual production has reshaped how we design, visualize, and shoot film environments. As a VFX MA student at UAL, I’ve always been curious about where the line between art and automation lies. When AI joins the scene—not as a background tool, but as an active collaborator in creative decision-making—what does that mean for human artists?
My critical research report explores the tension between automation and artistic agency in AI-assisted set dressing. While AI can speed up workflows, generate real-time environments, and even make aesthetic suggestions, it also challenges long-held ideas about authorship, originality, and creative control.
Applications of AI in Virtual Set Dressing
| Functional Module | Traditional Workflow | AI-Assisted Workflow | Benefits |
|---|---|---|---|
| Pre-visualization | Manually building models and storyboards | AI auto-generates concept sketches (e.g., Runway ML) | Faster prototyping, quick iteration |
| Lighting Layout | Manual lighting setups by technicians | AI adjusts lighting based on semantic context | Higher accuracy, real-time adaptability |
| Space & Prop Placement | Manually arranged by art department | AI generates spatial compositions via text prompts | Reduces labor, increases creative iteration |
| Scene Stylization | Concept artists + 3D integration | AI applies stylized render presets | Greater flexibility across design aesthetics |
The Big Question
My core question was this: Is AI helping artists, or replacing them? In practice, it’s not a simple binary. AI changes the nature of creative labour. Artists are no longer just makers; they become curators, directors, and supervisors of generative systems. This shift feels empowering on one hand—and precarious on the other.
To understand the impact, I looked at:
- How AI tools like Unreal Engine, Runway ML, and systems like FilmAgent are used in set design
- The philosophical debate over whether AI can be “creative”
- The ethical dilemmas: originality, authorship, and bias
- Industry perspectives and academic theories
Research Process
The writing process started with a review of creativity theories—like Guilford’s divergent thinking and Amabile’s componential model—emphasizing that creativity involves emotion, intention, and experience, all of which AI lacks.
From there, I explored current uses of AI in virtual production. Reports from IllusionXR, Voia, Raindance, and World Economic Forum showed how AI is already embedded in previsualization and set construction. I also read academic work like Xu et al.’s FilmAgent project, which uses large language models to simulate film crews.
Finally, I examined the risks: deepfakes, lack of consent, and algorithmic bias. Who’s really in control of the visuals? And are we designing with ethics in mind?
Table: Case Study Comparison – AI vs Artistic Control
| Case Study Name | Degree of Automation | Artistic Control | Originality Concerns | Reference Note |
|---|---|---|---|---|
| FilmAgent (Xu et al., 2025) | High | Medium | Yes | LLM-driven full-scene generation |
| Voia Virtual Tool | Medium | High | Clearly Controlled | Human-directed, AI-supported |
| Runway ML + UE Combo | High | Low–Medium | Potential Issues | Criticized for lack of transparency |
| LED Volume System | Medium | High | Project-Dependent | Integrates with director oversight |
Key Findings
Creative Shift
AI moves artists toward high-level roles—less manual work, more conceptual thinking. That sounds great, but it also reduces tactile authorship and increases dependence on opaque systems.
Control vs. Collaboration
AI can assist with layout, lighting, and composition, but results aren’t always predictable. This “black box” nature makes it harder to maintain full creative agency.
Ethical Uncertainty
The use of AI in storytelling raises difficult questions. If an AI reuses patterns from its training data, is it still original? Are we giving enough thought to authorship, bias, and consent?
What This Means for Us as Artists
This research reminded me that technological innovation doesn’t automatically equal artistic progress. We must stay critical. As artists, we need to ask:
- Am I using AI as a tool, or letting it shape my vision?
- Do I understand how the system works—and what it’s trained on?
- Am I protecting the ethics and integrity of my creative voice?
Like animation or any visual form, virtual production isn’t neutral. It’s a site of power, choice, and representation. We must design with intention, not just convenience.
Final Reflection
AI is here to stay in filmmaking—but it doesn’t have to replace us. Instead, it can extend our capabilities if we stay mindful, informed, and creatively brave. My biggest takeaway? Creativity in the age of automation isn’t about doing less—it’s about thinking more.