Source-linked AI summary
Designing Participatory AI: Creative Professionals' Worries and Expectations about Generative AI
Nanna Inie, Jeanette Falk, Steven Tanimoto
TL;DR
Creative professionals face speculation about generative AI’s effects, including possible threats to their work and reasons for curiosity. This paper surveys 23 professionals and uses thematic analysis to examine their views; responses varied in technical understanding and perceived AI agency. The authors propose four participatory-AI design foci and note that the preliminary findings come from a sample weighted toward software-oriented fields.
Problem
Generative AI may disrupt creative work, yet evidence about how creative professionals understand and expect to use it is limited in this paper’s framing.
Source: §1 · p. 1Method
The authors conducted an open-ended qualitative survey of 23 self-defined creative professionals and thematically analyzed their responses.
Source: §2 · p. 2, §2.1 · p. 2Results
Respondents’ descriptions of AI varied in technical depth and attributed agency; 10 responses paired superficial technical understanding with high computational agency, while none paired deep understanding with high agency.
Source: §3.1.1 · p. 2Takeaways & Limitations
The authors offer Understanding, Coping, Adapting, and Exploiting AI as categories for engaging creative professionals in participatory-AI design.
Source: §4 · p. 4Most respondents came from software-oriented creative domains, so their views may be more open and welcoming toward AI; the authors call for broader field representation and richer interview data.
Source: §2.1 · p. 2, §5 · p. 5
Abstract
from arXiv · showhide
Generative AI, i.e., the group of technologies that automatically generate visual or written content based on text prompts, has undergone a leap in complexity and become widely available within just a few years. Such technologies potentially introduce a massive disruption to creative fields. This paper presents the results of a qualitative survey ($N$ = 23) investigating how creative professionals think about generative AI. The results show that the advancement of these AI models prompts important reflections on what defines creativity and how creatives imagine using AI to support their workflows. Based on these reflections, we discuss how we might design \textit{participatory AI} in the domain of creative expertise with the goal of empowering creative professionals in their present and future coexistence with AI.
1 INTRODUCTION AND BACKGROUND
Generative AI raises concerns about creative professionals’ ownership and agency while also attracting curiosity about its potential. The paper categorizes creatives’ views and considers participatory AI as a way to inform their coexistence with these systems.
1 INTRODUCTION AND BACKGROUND
- Generative AI is viewed both as a possible threat to creative ownership and agency and as a technology that some creatives approach with curiosity and excitement.Potential threats include generating high-quality content, expanding content variety, and personalizing content for users.Source: §1 · p. 1
- Participatory AI aims to include wider publics in AI development and deployment, with values of inclusion, plurality, collective safety, and ownership.The idea draws on participatory design’s aim of empowering workers to codetermine information systems and their workplaces.Source: §1 · p. 1
- The paper surfaces and categorizes concerns and expectations about generative AI from creatives of different types, as an initial empirical inquiry into designing participatory AI.The authors identify the work as their first empirical research into how participatory AI might be designed and performed.Source: §1 · p. 1
- Prior work examined expectations for AI tools among creative writers and designers, while this paper addresses creatives’ concerns and expectations across different types.The cited examples include creative writers, game-level editors, and game designers.Source: §1 · p. 1
- The paper proposes four participatory-AI design foci—Understanding, Coping, Adapting, and Exploiting AI—alongside new conceptions of creativity and categories of concern and excitement.The authors present these findings as preliminary topics that may inform participatory design so creatives can influence technologies that change their work practices.Source: §1 · pp. 1–2
2 METHODS
The authors used an asynchronous qualitative survey and thematic analysis to examine creative professionals’ understandings of and attitudes toward AI and creativity. Its 23 respondents came from varied fields, with most working in software-oriented creative domains.
2 METHODS
- An asynchronous qualitative survey with open-ended questions was circulated through the authors’ networks and social media over approximately two months in late 2022.The term “creative” was left to respondents’ self-definition, and the survey was approved by the authors’ universities’ ethics committees.Source: §2 · p. 2
2.1 Participants
- The 23 respondents lived in six countries and worked across varied fields, but most came from software-oriented creative domains.The authors caution that the findings should be read with this participant-profile limitation in mind.Source: §2.1 · p. 2
2.2 Survey and analysis
- The survey combined demographic questions with six questions designed to elicit general understandings of and attitudes toward AI and creativity.Questions about excitement and worry were swapped for approximately half the respondents to avoid priming in a particular direction.Source: §2.2 · p. 2
- The authors conducted thematic analysis by coding responses individually and clustering codes into sub-themes.The analysis followed Braun and Clarke.Source: §2.2 · p. 2
3 SURVEY RESPONSES
Respondents differed in how technically they understood AI and how much agency they attributed to it, while their views of AI creativity ranged from input recombination to requirements such as intention, sentience, and self-awareness. They described reasons both to worry and not to worry about generative AI, alongside ways it might support creative work.
3.1 How intelligent or creative is generative AI?
- Respondents ranged from superficial to deep technical understandings of AI and from attributing it no agency to attributing it high computational agency.Ten responses paired superficial technical understanding with high agency; none paired deep understanding with high agency.Source: §3.1.1 · p. 2
3.1.2 New definitions of creativity.
- Some respondents treated recombination of existing inputs as potentially creative, while others required originality, agency, intention, sentience, or self-awareness.The reflections prompted varied conceptions of creativity, without implying that these should become a unanimous scholarly definition.Source: §3.1.2 · p. 3
3.2 I Am Not Worried (Yet)
- Nine respondents reported no worry and six no worry yet; three unequivocally worried, while copyright concerns were directly mentioned by only one respondent.The authors note that this may mean the issue was not seen as threatening by surveyed creatives or that they expected technological solutions.Source: §3.2 · pp. 3–4
- Other concerns included increased output diluting quality or genres, and AI bypassing exploration and potentially weakening creative processes.One respondent also raised questions about whether computational processes taking over parts of human thought might affect cognition.Source: §3.2 · p. 3
- Reasons for not worrying included dependence on human input, unconconvincing or samey output, and creative work too complex for AI to imitate.Respondents also described human verification, problem solving, and client communication as important parts of creative work.Source: §3.2 · p. 4
3.3 Exciting Times Ahead!
- Thirteen respondents were unequivocally excited about AI in creative work; they imagined it increasing productivity and providing inspiration through larger solution spaces.Proposed uses included automating repetitive tasks, streamlining processes, and rapidly trying different ideas.Source: §3.3 · p. 4
- Some respondents expected AI to improve output through collaboration and to support convergent work such as decision-making and evaluation.The respondents did not describe specific ways for evaluation to occur.Source: §3.3 · p. 4
4 DISCUSSION: OPPORTUNITIES FOR PARTICIPATION
The discussion frames participatory AI around helping creatives understand, cope with, adapt to, and exploit AI. It argues that meaningful participation can help integrate AI into creative work and reveal new uses.
4 DISCUSSION: OPPORTUNITIES FOR PARTICIPATION
- The four design foci are understanding, coping with, adapting to, and exploiting AI, aligned with what end-users know, feel, do, and dream.The authors propose these categories as a framework for engaging professional creatives in participatory AI design.Source: §4 · p. 4
4.1 Understanding AI
- A truthful understanding of AI and its limitations is presented as a first step toward co-creation, with developers sharing responsibility for accessible and explainable systems.The discussion notes that no responses combining deep technical understanding with high ascribed agency were observed.Source: §4.1 · p. 5
4.2 Coping with AI
- Coping includes recognizing human properties of creativity, sharing worries and strategies such as avoiding AI, and considering how AI might acknowledge its data sources.The authors suggest creatives hone their skills in creating and evaluating creativity.Source: §4.2 · p. 5
4.3 Adapting to AI
- Adapting involves including creatives in model development, while developers can tailor systems to specific professions and crafts, including automating repetitive or boring work.The discussion distinguishes adaptation from coping as more comprehensive inclusion in developing specific generative AI models.Source: §4.3 · p. 5
4.4 Exploiting AI
- Exploiting AI may involve new creative tools, media, and jobs, while participation could improve integration into work practices and inspire new uses.The authors hypothesize that creative professionals' participation may help leverage their competencies to imagine new avenues for the technology.Source: §4.4 · p. 5
5 CONCLUSION AND FUTURE WORK
The authors encourage involving creatives in generative AI development and considering how they can understand, cope with, adapt to, and exploit AI. They caution that the study's sample and survey method limit the scope of its preliminary findings.
5 CONCLUSION AND FUTURE WORK
- The authors encourage involving creatives in technology development and larger project issues, and orienting participatory AI around understanding, coping with, adapting to, and exploiting AI.They describe the findings as insights into how creative professionals anticipate effects on their work practices.Source: §5 · p. 5
- Because most represented fields are software-oriented, respondents' views may be more open and welcoming toward AI.The authors call for more evenly distributed representation across creative fields and richer data from interviews.Source: §5 · p. 5
- Respondents came from different creative industries, whose everyday work may not be affected by generative AI in the same ways.The authors did not characterize how individuals' understanding of AI relates to their worries or expectations, citing the need for a larger group and deeper investigation.Source: §5 · p. 5
- Future work could categorize industries and identify which specific work tasks may be affected, as well as examine ways to support those practices with AI.Source: §5 · pp. 5–6
A RESPONDENTS OVERVIEW
The section provides an overview of survey respondents.
A RESPONDENTS OVERVIEW
- The table header names respondent number, age, country, gender, and a prompt about profession and creativity at work.Source: A RESPONDENTS OVERVIEW · p. 7
- Table 1 is identified as an overview of survey respondents.Source: Table 1 · p. 7
B GRAPH OF RESPONSES TO “WHAT IS AI?”
Figure 1 depicts the distribution of answers by technical depth and ascribed AI agency.
B GRAPH OF RESPONSES TO “WHAT IS AI?”
- Figure 1 shows the distribution of answers by technical depth and ascribed agency of AI.Source: Figure 1 · p. 8