Zixuan Li

Introduction

Aim and research questions

Prototype development process

Secondary research

Better ways to go brainstorming:

• Positive feedback: participants being subjected to positive feedback in brainstorming significantly improves the efficiency. (Yoshida et al., 2016).
• Harmonious environment: creating a harmonious brainstorming environment without evaluating the merits of ideas in the process. (Wilson, 2016)
• Trained leader: successful brainstorming requires a trained leader. (Wilson, 2006)
• Setting a target: setting a target for the number of ideas for each brainstorming that is two times higher than one’s ‘typical performance’ can significantly improve idea generation. (Wilson, 2006)
• Board game: designing the brainstorming process to resemble a board game can significantly increase participant engagement, as well as stimulate more innovative ideas. (Alenquer & Gan, 2011)
• Role-playing: the use of situational simulation and role-playing in brainstorming allows participants to think about the problem from the user’s point of view, which helps to increase the depth of empathy and creativity. (Alenquer & Gan, 2011)

Survey

A questionnaire was distributed to students studying design at Loughborough University who had experience of brainstorming to investigate their brainstorming preferences and attitudes towards AI.

Designer preference：

3 possible ways

As a result, three possible ways for designers to use AI were identified to try and see which would be more appropriate for designers’ design patterns

So,using different design methods combined with AI is the most appropriate and stable way to assist designers in brainstorming.

Then, the main task was to explore which design pattern combined with AI would be more stimulating for designers to perform better in brainstorming.So I chose the way of the two most recognised, scamper and 6hats.

Final prototype

I used two brainstorming questions for the crossover experiment because the same person can’t brainstorm a question twice.

Evaluation methodology

Quantitative analysis

Data from all the above quantitative analyses will be analysed through SPSS tools.

Qualitative analysis

After the participants had completed all the prototype tests, a short interview was conducted to ask them about the strengths and weaknesses of SCAMPER and 6HATS after working with the AI, their tendency to use it, what needs to be improved, and what other ideas they have for AI involvement in brainstorming.

Results of the final evaluation

Quantitative analysis

In summary, they felt better subjectively about scamper, but scamper didn’t make a significant difference to the quantity or quality of their IDEA.

Qualitative analysis

In the qualitative analyses, the data were all derived from the participants’ exit interviews
and each question covered a different perspective and scope

Advantages & Disadvantages

Preference

Improvement

• 55% of the participants asked the AI to provide relevant information and similar cases.
• 25% of the subjects asked the AI to be more streamlined and visualized.
• 25% asked the AI to give them feedback and evaluation of their ideas.
• 15% asked the AI to keep exploring and iterating and refining based on their existing ideas.
• 10% asked the AI to provide images that inspire

Prototype improvements

According to the analyses, although there was no significant difference in the quantity and quality of ideas generated during the brainstorming process using the SCAMPER and the Six Hats methods, from the subjective point of view of the designers, the SCAMPER was superior to the Six Hats in terms of satisfaction, perceived quality of the ideas, and quantity of ideas. And based on the results of the qualitative analysis, I upgraded the SCAMPER prototype and conducted a small test.

• AI provides examples, knowledge and pictures in addition to guiding questions.
• The interface is more visual and logical (Mindmap).

Analysis

Guidelines

This paper investigates the impact of guided questions generated by AI in combination with
different design practices (SCAMPER & 6HATS) on the brainstorming process of designers,
and a series of conclusions are drawn. These conclusions can generate some guidance on
how to better integrate AI into the designer’s inspiration phase in the future.

1. SCAMPER is more suitable for the ideation stage of brainstorming compared to 6 Hats.
   When combined with AI, the guiding questions generated by SCAMPER align more with
   the thought patterns and logic of most designers, thereby achieving higher satisfaction.
   26
2. 6 Hats is better suited for the research phase or design review stages. The 6 Hats
   framework is more practical but less inspirational. During the actual development phase,
   users should have the option to choose between different models.
3. AI should provide a plethora of examples for each dimension of guiding questions.
   Textual case descriptions can better stimulate designers’ thinking, including both directly
   relevant and similar cases.
4. AI should also provide a vast amount of relevant knowledge and tips for each
   dimension of guiding questions. Due to cognitive limitations and knowledge gaps,
   designers can easily stagnate; substantial input of background knowledge can help
   inspire them more effectively. However, care must be taken that these tips are not
   overly specific, as they may constrain thinking.
5. Use images cautiously during the brainstorming process. Images generated by AI often
   lack practical considerations and overly specific images can restrict designers, making it
   difficult for them to think outside the box.
6. In the AI‐assisted brainstorming process, more integration and visualization are
   needed. Designers generally prefer to see all guiding questions organized by different
   dimensions using mind maps or tables, which helps them clarify their thoughts.