Where Are We So Far? Understanding Data Storytelling Tools from the Perspective of Human-AI Collaboration

Title of the Paper

Where Are We So Far? Understanding Data Storytelling Tools from the Perspective of Human-AI Collaboration

Paper Information

• Subject Area: Data storytelling tools and human-AI collaboration
• Keywords: Data storytelling, human-AI collaboration, artificial intelligence, visualization, design patterns, data stories, automated generation, story planning, role distribution, human-computer interaction

Research Background and Issues

• Identified Problems or Challenges:

  • Data storytelling requires creators to possess diverse skills and invest significant time and effort.
  • Designing effective human-AI collaboration tools faces numerous challenges, such as balancing system automation with human control and fostering communication and collaboration between humans and AI.
  • While existing studies have outlined the characteristics or technical applications of data storytelling tools, there is a lack of a systematic perspective on how humans and AI collaborate within these tools.

• Significance:

  • Data storytelling effectively conveys data insights and plays a critical role in fields such as data science, business analytics, and academic research.
  • Combining the strengths of artificial intelligence and human creators to design efficient collaborative tools can significantly enhance data processing and storytelling efficiency.

• Research Motivation and Related Work:

  • Aims to systematically examine how data storytelling tools support human-AI collaboration and identify their design patterns and future research directions.
  • Reviewed existing literature and analyzed tool design from two perspectives: story process stages and human-AI collaboration roles.

Proposed Solution

• Proposed Methods or Solutions:

  • Developed a framework to analyze tool design from two dimensions: (1) stages of data storytelling—analysis, planning, implementation, and communication; (2) roles played by humans and AI at each stage—creator, assistant, optimizer, reviewer.
  • Collected and coded existing data storytelling tools to observe the stages covered and collaboration patterns.

• Innovations:

  • The first systematic evaluation of human-AI collaborative data storytelling tools from the perspectives of roles and stages.
  • Identified common collaboration patterns, expanding research on creative support tool design.
  • Provided research opportunities based on tool design, suggesting ways to maximize automation and human control.

• Implementation Steps:

  • Created a literature dataset, filtering papers describing data storytelling tools related to human-AI collaboration.
  • Coded the stages and collaboration role patterns covered by the tools, comparing design patterns and application effects.
  • Offered an interactive tool browser for the academic community to continuously update.

Research Outcomes

• Specific Results:

  • Compiled 60 papers on data storytelling tools published between 2010 and 2023.
  • Identified two most common collaboration patterns: (1) human creator + AI assistant; (2) AI creator + human optimizer.
  • Found that the design of data storytelling tools has gradually transitioned from fully manual to human-AI collaboration, emphasizing human control over story content.

• Advantages:

  • Summarized design patterns and collaboration challenges through systematic analysis.
  • Demonstrated complex design approaches that adapt to human needs and leverage AI strengths.
  • Provided recommendations for future tool design, such as using generative AI to enhance collaboration experiences.

• Experimental or Evaluation Results:

  • Coverage of data storytelling tools: the implementation stage has been widely studied, while the analysis and communication stages have received less attention.
  • Trends in story formats: shifting from single static visualizations to complex and diverse animated stories.
  • Tool designs are increasingly supporting comprehensive, cross-stage, and hybrid collaboration approaches.

• Limitations and Future Directions:

  • Limitations: The current framework lacks detailed granularity and does not deeply explore task-level role distribution; it does not cover tools driven by the latest generative AI technologies.
  • Future Directions: Suggests studying task-level collaboration patterns, investigating tool performance across different communication channels, and expanding the framework's applicability to include visualization creation and creative support tools.

The above content summarizes the core information of the paper, helping academic researchers understand the background, contributions, and future research directions of the study.