Catalyst for Creativity or a Hollow Trend?: A Cross-Level Perspective on The Role of Generative AI in Design

Research Background and Issues

• Identified Problems or Challenges:

  • The transformative potential of generative AI image tools in design education and practice has sparked significant debates around creativity and ownership.
  • With the rise of generative AI tools (e.g., MidJourney, DALL.E) in 2022, designers' learning experiences have been divided into "pre- and post-tool emergence," leading to a divergence in value systems among designers.
  • Professional designers are concerned about the potential loss of traditional creativity and design skills, while novice designers are more inclined to embrace the technology.

• Why It Matters:

  • Generative AI lowers the technical barriers to design, enabling more individuals to engage in high-quality visual creation.
  • The design industry faces challenges in redefining creativity, addressing disputes over ownership of creations, and integrating such technologies into design education.

• Research Motivation and Related Work:

  • Investigating prior models of creativity, such as Sawyer's eight characteristics of creativity framework, to explore the impact of generative AI on these traits.
  • Positioning generative AI as a potential "dual-use technology" with both significant positive potential and possible risks or negative impacts.

Solution

• Proposed Methods or Solutions:

  • Conducting 28 design-instance-based interviews with designers of varying experience levels (beginners, senior design students, and professional designers) to explore the impact of generative AI on design education and practice.
  • Comparing differences in design processes, skill definitions, and ownership evaluations across groups with different levels of experience.

• Innovative Contributions:

  • Cross-Level Comparison: The study is the first to compare the impact of generative AI on design learning and practice across different experience levels.
  • Instance-Based Interviews: By having participants reflect on specific AI-generated design works, the study reveals practical challenges and support points in using the technology.
  • Dual-Use Technology Framework: Comparing generative AI to historical dual-use technologies, the study proposes policy recommendations.

• Implementation Steps and Key Techniques:

  • Recruit participants across experience levels (beginners, students, and professionals) and collect design instances with AI-generated backgrounds.
  • Conduct interviews to gather detailed information on the impact of generative AI, particularly regarding creative processes, education, skill descriptions, and ownership.
  • Perform thematic analysis based on the Braun and Clarke method to extract core research findings.

Research Findings

• Specific Findings:

  • The impact of generative AI on design roles is summarized as a "value system split." Beginners tend to view AI as a tool for enhancing efficiency and creation, while advanced design students and professionals are more concerned about its disruption to traditional skills and creativity.
  • AI is inevitably reshaping the design industry, pushing education and practice toward a hybrid skill model that balances traditional design and AI tool proficiency.
  • The hidden labor and practical challenges of using AI (e.g., prompt engineering, cultural biases, inconsistencies) prompt researchers to question the long-term practical value of AI.

• Advantages Over Existing Solutions:

  • Unique Perspective: The study categorizes and compares the impact of AI on the design ecosystem based on experience levels, rather than focusing on a single group.
  • Revealing Hidden Challenges: Through instance-based interviews, the study uncovers complexities and practical issues encountered by users (e.g., editing difficulties, inconsistent outputs).

• Experimental or Evaluation Results:

  • Beginners generally perceive AI as a core tool for the future of design, appreciating its high applicability and quality, while advanced students and professionals exhibit a split in discourse on the value of skills.
  • AI tools often lack explainability, requiring designers to engage in trial-and-error processes, which increases workload.
  • Participants highlighted severe cultural and linguistic biases in AI tools, which could lead to misrepresentation in design outcomes.

• Limitations and Future Directions:

  • Limitations:
    • The sample is concentrated in the United States, which may limit the global applicability of the findings.
    • The rapid evolution of tools may render the research findings quickly outdated.
    • The study focuses only on prompt-based tools and does not cover other generative methods (e.g., sketch-to-image).
  • Future Directions:
    • Explore the specific forms and timing for integrating generative AI-assisted tools into design education.
    • Propose regulatory and policy frameworks for design AI, focusing on ethics and ownership issues.
    • Develop AI support tools with strong explainability to reduce feedback loop challenges between designers and technology.

This study provides critical insights into the future development of design education and practice, while highlighting the complex interplay between technological empowerment and the preservation of traditional skills. Addressing key technical challenges could further influence the transformation of the design industry.