StoryVerse is a world-building tool for writers that offers dynamic exploration of story elements through maps, timelines, and more.
It provides a scaffolding for creativity, thoughtfully integrating AI to illuminate connections across notes, research, inspiration, and ideas.
HMW: How might we enhance the world-building activities of novelists in ways that support their unique processes?
Background:This project was done over an 8 month capstone sequence as part of my MHCI+D at the University of Washington. At the onset of the project we were directed by our program directors to. After a lengthy exploration process our team focused on helping creative fiction writers with a focus on AI tools given our team’s interest in emerging technology and stories as a medium.
Other Members (Significant* Roles in addition to UX Research/Design ): Carver: Storytelling+boards Presentations/Videography Nat: Visual/UI Design and StoryBoard Artworks Tavin: Visual/UI Design
Advised By: Michael Smith
Our research explored fiction writers’ attitudes and perceptions around generative AI. We found that writers mostly do not want AI to write for them, but some are open to other use cases. We also learned about their rigorous note-taking, research, and world-building processes. We saw an opportunity to design a tool to support writers in the visualization and retrieval of information needed to create the worlds of their stories while also rethinking the possible future of AI’s role in creative writing.
Research
The first 3 months of our project consisted of undertaking a research study on the perceptions, thoughts, and uses of generative AI among creative fiction writers.
The study involved interviews with a diverse group of seven published writers, including two subject matter experts (SMEs).
The findings reveal a nuanced landscape of writers' perceptions of AI in writing. While many participants expressed concerns about AI-generated content flooding the market and exacerbating the problem of overcrowding in the publishing industry, others saw opportunities for experimentation and novel ways to support their writing practices.
Why AI? We witnessed the initial boom and out of the blue emergence of ChatGPT during our program.
ChatGPT has 1.8 billion visitors per month. Generative AI has the potential to automate 26% of work tasks in the arts, design, entertainment, and media sectors. The Writers Guild of America strike shed new light into the potential dynamics of AI in the creative industry.Artificial intelligence (AI) has gained significant attention in creative fields in recent years; specifically, its capabilities to produce creative works have raised foundational questions about meaning [1], ownership [2], and authenticity [3].
Generative AI tools can produce text that mimics human writing to a remarkable degree [4]. Moreover, these tools have the capabilities to produce content in ways that impact the overall process of narrative creation [5].
Initial Desk Research Findings:
1.
The scale and potential impacts of AI are unprecedented:
Reportedly, generative AI has the potential to automate 26% of creative work tasks in the arts, design, entertainment, and media sectors. [12]
2.
The volume and ubiquity of AI-generated content are already impacting the industry:
Amazon has seen a boom in AI-generated books sparking concerns about copyright and threats of job replacement [13]
News such as the Writers Guild Association strike has begun to explore what guardrails might entail in defining the relationship between human writers and AI tools [16].
3.
There are emerging use cases for AI tools
Constraints, antagonism, and better alignment with intent through AI tools could arguably better serve writers than popular counterparts today. [3][17]
Some writers with technical literacy have been building their own tools to achieve vastly different results from mainstream tools. [18]
Mental models such as hip-hop sampling as an emergent practice are being embraced by writers experimenting with AI tools. [19]
Methods
Our research study consisted of 7 semi-structured participant interviews and 2 subject matter expert interviews. A breakdown of our participant profile can be found in Table 1 below.
Research Question(s):
What are the attitudes of creative fiction writers towards generative AI tools in relationship to the writing process?
This inquiry manifested through three sub-questions reflecting what we wanted to learn:
How do fiction writers currently write from start to finish? How do they feel about the role of AI in creative writing? How does external feedback or collaboration factor into their process, if at all?
Interview Structure and Objectives
Most of our research focused on qualitative interviewing, with some participants sharing tools, work, or online resources integral to their writing process. The central part of the interview was a 40-minute section split among our three sub-questions. Given that we primarily focused on the attitudes and perspectives of creative fiction writers, we primarily conducted qualitative research through semi-structured conversational interviews.
To get a deeper perspective on our participants' attitudes and mental models of AI, we curated various quotes regarding current opinions and use cases from other writers. After letting the participant read each quote, we asked whether or not they changed or reinforced anything about their understanding or opinion of generative AI tools (shown in Figure 2 in the next page)
Results
1.
Most participants view AI-produced writing as formulaic, and therefore found it not worth incorporating directly into their works
“I've been working with [AI] for a long time, and there just is no AI that could produce a plot you want [...] you would ask it to come up with a plot and it would be the dumbest thing you ever heard in your life.” — P1, Columnist & Novelist (AI Novella Author)
2.
Participants largely don’t want AI to write for them; they’re choosing to write out of passion, and even writer’s block can be a valuable part of the process.
"It does seem like if you are using an AI to do the writing for you, do you even like writing? And if you don't, why are you writing?" — P2, Speculative sci-fi writer
3.
Most participants see getting their books in the hands of interested readers as the salient problem in the experience of being a professional writer.
4.
The human experience enriches feedback, particularly because of people’s capacity to contextually interpret the work.
5.
Writers’ attitudes on AI impacting the industry have been partly defined by concern about potentially worsening problems of excess volume in publishing.
Payouts for that month to authors were lower...It was because so many books were added to Amazon that were written by chatbots...They use the bots to write the books, and then they just uploaded the books basically with stupid covers. This is what they think happened.— P6, Romance Author
6.
Some fiction writers have started developing personalized use cases for AI out of interest in experimentation and discovering novel ways to support their writing.
There's this interesting idea...which is like the model as an antagonist. You have the model do something, and then you're like, well, I definitely shouldn't do that...I have to do something more creative, more innovative. — SME2, Researcher
Finding #5:
Current solutions disproportionately impact new or foreign authors producing legitimate work. These are not communities we should be leaving behind for the sake of convenience. — SME1, Publication editor
Finding #6:
“A friend of mine trained AI on his own writing, and he’s using it when he needs ideas based on his own work so he considers it a memory aid.” — SME1, Editor
Design Principles and Opportunities
Based on the research outcomes, four design principles were identified as crucial for any design intervention in this space:
The report highlights three significant design opportunities for addressing the needs and pain points of fiction writers:
1. The first opportunity involves developing tools that enhance supporting activities in writing, such as researching, organizing, and world building. These tools can utilize technology, including AI, to create personalized databases and networks of information that inform writers' work.
2. The second opportunity focuses on connecting novelists with writing partners who share similar interests and can provide meaningful feedback. By facilitating these connections, writers can receive feedback that contextualizes their work and helps them refine their craft.
3. The third opportunity centers around assisting writers in marketing and promoting their work to relevant audiences. Writers often face challenges in effectively marketing their books, particularly in non-big publishing spaces. Design interventions could provide resources, educational materials, and platforms that empower writers to navigate the complexities of book marketing and reach their intended audiences.
1. How might we develop novel ways to enhance supporting activities for creative fiction/novel writing (i.e. researching, organizing, worldbuilding, etc.)? This design opportunity is informed by the following research findings: • Finding #2: Participants largely don’t want AI to write for them; they’re choosing to write out of passion, and even writer’s block can be a valuable part of the process. • Finding #6: Some fiction writers have started developing personalized use cases for AI out of interest in experimentation and discovering novel ways to support their writing.
2. How might we connect novelists with writing partners who share similar interests and can provide meaningful feedback? This design opportunity is informed by the following research findings: • Finding #4: The human experience enriches feedback, particularly because of people’s capacity to contextually interpret the work.
3. How might we help writers market and promote their writing to relevant audiences?
This design opportunity is informed by the following research finding: • Finding #3: Most participants see getting their books in the hands of interested readers as the salient problem in the experience of being a professional writer.
• Finding #5: Writers’ attitudes on AI impacting the industry have been partly defined by concern about potentially worsening problems of excess volume in publishing.
Ideation
Our team started off by coming with a wide range of ideas as potential design responses to our 3 design opportunity HMWs. Each of us came up with 20 ideas for a total of 80 responses. (Insert Image of 80 Ideas)
We created a showcase of top 3 potential responses:
Selected Idea:
In the end we chose to go with the first design opportunity:
How might we develop novel ways to enhance supporting activities for creative fiction/novel writing (i.e. researching, organizing, worldbuilding, etc.)?
There were several reasons we went with the above. As a field, HCI seemed very well positioned to be in a supporting role for writers. Computers specifically can help augment and extend a writer's vision and this can serve great value for audiences down the line as we see in many new media fields with VFX, interactive games, digital worlds, etc.
Finding feedback partners was not core to the writer’s act of writing and marketing and promotion to audiences was post work activity to help a written piece gain reach not* enhance or fundamentally change the way it manifests itself.
Prototyping
With the writers we talked with there was a pattern of explanations of complex and arduous world building processes that are rigorous and thorough with conceptions, details, and hopes.
Our selected idea was a world building aid: a map and design system for writers to imagine the world they want to bring to life.
Like a story universe builder for writers and thus we decided to call it: “StoryVerse”
Design Concept We envisioned StoryVerse to do 3 things:
A memory aid for inspiration and research details in an easy to navigate system.
Direct the consistency and the interplay of detailed elements
Envision the evolution of a story in context
To envision the key/signature interactions we needed to create for the interaction model we created an initial Storyboard:
Feedback Session We gathered some initial feedback from our instructors who pointed out the drawbacks of just having a note management system compared to our initial worldbuilding vision. As a result we decided to gather a bit of feedback from a writer we interviewed earlier.
WhiteBoarding With the writer’s feedback we went to the drawing board to visualize the key interfaces for the core interactions we saw as needed for our original vision.
UI Design V1 Our initial designs brought our whiteboard ideas to life.
Design Changes The initial UI Elements were in place, but one thing we realized was the need for a central home for a user’s projects.
UI Design V2 The second version of our design introduced a fundamental change in how users navigate the interface and project pages to mimic more of a workspace interface rather than a simple website with multiple media functionalities. The navigation of a project was made more clear, and users can have multiple function pages open: Notecards, Timelines, Maps, etc.
