Understanding Screenwriters' Practices, Attitudes, and Future Expectations in Human-AI Co-Creation

Research Background and Issues

• What problems or challenges did the authors identify?

  1. Common challenges faced by screenwriters: including lack of inspiration, insufficient coherence in narrative texts, lack of content depth, and inadequate emotional resonance.
  2. The integration of AI into various workflows of screenwriting has not yet been fully developed, remaining limited to certain isolated stages without covering the complete creative process.
  3. Although AI generation technologies help transition from information organization to content creation, they lack emotional understanding capabilities, and the generated content often lacks nuanced emotions and cultural diversity.
  4. Future expectations for AI collaboration tools have not been thoroughly studied, especially regarding improvements in full-process collaboration and content customization.

• Why is this issue important?
  As AI's role in creative fields continues to rise, its potential for collaboration with humans is significant in enhancing creators' productivity, inspiring creativity, and even transforming traditional industry workflows. At the same time, AI integration will inevitably spark new debates about copyright, job displacement, and authorship, bringing profound impacts to the industry.

• Research Motivation and Related Work
  Although tools like ChatGPT and AI Dungeon have been used in creative writing and screenwriting, their applications are mainly focused on single tasks and have not deeply explored every stage of the workflow. Furthermore, much of the current exploration centers on the text generation capabilities of large-scale language models (LLMs), neglecting the potential for generating visual and dynamic elements. This study aims to fill this research gap by interviewing screenwriters to understand their current practices, attitudes toward AI, and future expectations.

Solutions

• What methods or solutions did the authors propose?

  1. Conducted semi-structured interviews with 23 screenwriters to explore how AI is applied in specific screenwriting workflows.
  2. Identified the main application areas and challenges of AI in screenwriting and summarized six key stages of the screenwriting workflow: goal and ideation, synopsis and outline, character design, story structure and plot, dialogue creation, and script text.
  3. Categorized interview results to summarize screenwriters' expectations for AI tools in four roles: "Actor" (simulating characters), "Audience" (providing feedback), "Expert" (offering advice), and "Executor" (completing tasks as needed).

• What are the innovations of this solution?

  1. Specifically breaking down the collaboration between screenwriters and AI into actual workflow stages, providing targeted improvement suggestions.
  2. Proposing four role classifications for AI's future functionalities and clarifying their potential application scenarios in different stages of screen creation.
  3. For the first time, approaching the research from the perspective of screenwriters' "needs" and "expectations," quantifying and refining their feedback on AI usage and potential improvement directions.

• What are the implementation steps and key technologies used?

  1. Semi-structured interviews: Designed 27 open-ended questions covering basic workflows, experiences with AI tools, and future needs.
  2. Analysis: Recorded 1992 minutes of audio from interviews, generating 558 pages of transcription, and applied open coding methods using a six-step thematic analysis approach for classification.
  3. Data collection and classification: Extracted stage-specific characteristics of workflows and screenwriters' feedback on AI tools through transcript analysis, summarizing functional advantages, disadvantages, and application prospects.

Research Findings

• What specific findings were achieved?

  1. Current Practices: Most participants (78%) have already applied AI tools in their screenwriting work, particularly in story structure and plot development, script text generation, goal and ideation generation, and dialogue creation.
  2. Attitude Analysis: Participants expressed multi-faceted attitudes toward AI, including positive views on its ability to quickly generate ideas and reduce trial-and-error costs, but negative feedback on its poor accuracy and lack of emotional understanding.
  3. Future Needs: Screenwriters explicitly expressed expectations for AI to play the roles of "Actor," "Audience," "Expert," and "Executor," and pointed out their application scenarios in different screenwriting stages.

• What advantages does it have compared to existing solutions?

  1. Covers the entire screen creation workflow and refines the tasks and effects AI can intervene in at each workflow stage.
  2. Proposes specific application needs for AI in intelligent role simulation, enhancing plot logic, and optimizing visual expression.
  3. Emphasizes the design possibilities of personalization and multimodal interaction (text, images, sound, and even dynamic video), filling gaps in existing solutions in this area.

• What are the experimental or evaluation results?
  The current most effective stages for AI usage are concentrated in creative generation and simple character design. Participants generally reported limited effectiveness in complex plot generation and dialogue creation. AI needs to improve its accuracy, understanding of multi-dimensional emotions, and ability to perceive complex scenarios to meet user needs.

• Limitations and Future Directions

  1. Cultural and Regional Limitations: All participants were from China, which may introduce cultural biases; further validation is needed for applicability to the global screenwriting industry.
  2. Breadth of Experience: Participants' proficiency with AI tools varied, potentially leading to differences in insights; future research could stratify findings based on proficiency levels.
  3. Interface Design and Technical Implementation: Specific AI prototype tools for screen creation have not yet been developed; future research could focus on multimodal interaction design and optimization of human intervention.
  4. Extended Research Directions: Explore how AI further impacts screenwriters' professional identity, authorship rights, and the boundaries between human creativity and machine intelligence.