Source-linked AI summary
User Experience Design Professionals' Perceptions of Generative Artificial Intelligence
Jie Li, Hancheng Cao, Laura Lin, Youyang Hou, Ruihao Zhu, Abdallah El Ali
TL;DR
GenAI’s effects on UXD raise questions about whether it will assist designers or intensify harms such as job displacement and skill degradation. The paper addresses these questions through interviews with 20 UX Designers and finds that experienced designers view GenAI as assistive while emphasizing human agency and serious risks for junior designers. It also identifies design-handoff friction as a practical boundary relevant to responsible human–AI collaboration.
Problem
The paper examines how GenAI affects UXD practice and whether concerns about its consequences are warranted.
Method
The researchers conducted in-depth interviews with 20 UX Designers of diverse experience working across companies of different sizes.
Results
Experienced designers report confidence in originality, creativity, and empathy, view GenAI as assistive, and retain human agency while expressing serious concerns about junior designers’ skill degradation and unemployment.
Takeaways & Limitations
Responsible UXD collaboration should balance AI efficiency with human creativity, agency, and attention to copyright, ownership, and access.
Takeaways & Limitations
The study is bounded by frictional handoff between UX Designers and development teams in current UX practice.
Abstract
from arXiv · showhide
Among creative professionals, Generative Artificial Intelligence (GenAI) has sparked excitement over its capabilities and fear over unanticipated consequences. How does GenAI impact User Experience Design (UXD) practice, and are fears warranted? We interviewed 20 UX Designers, with diverse experience and across companies (startups to large enterprises). We probed them to characterize their practices, and sample their attitudes, concerns, and expectations. We found that experienced designers are confident in their originality, creativity, and empathic skills, and find GenAI's role as assistive. They emphasized the unique human factors of "enjoyment" and "agency", where humans remain the arbiters of "AI alignment". However, skill degradation, job replacement, and creativity exhaustion can adversely impact junior designers. We discuss implications for human-GenAI collaboration, specifically copyright and ownership, human creativity and agency, and AI literacy and access. Through the lens of responsible and participatory AI, we contribute a deeper understanding of GenAI fears and opportunities for UXD.
1 INTRODUCTION
GenAI expands creative and productive possibilities while raising concerns about unintended consequences, homogenization, job loss, and intellectual-property harms. The paper examines how these opportunities and risks bear on UXD, a user-centered practice requiring creativity, collaboration, problem solving, sense-making, and empathy.
- GenAI landscape: GenAI tools generate multimodal content from natural-language prompts, including text, images, video, and audio.Examples include ChatGPT, DALL-E 2, Stable Diffusion, and Midjourney.
- Opportunities and concerns: GenAI adoption has generated excitement about rapid advances alongside anxiety about unanticipated consequences and model weaknesses propagating downstream.The paper also notes concerns about homogenization.
- Opportunities and concerns: Creative professionals face potential reputation damage, economic and job loss, plagiarism, copyright infringement, and reduced human involvement.These concerns motivate attention to creative professionals’ knowledge and interests in responsible AI development.
- UXD focus: The paper investigates GenAI’s impact on UXD practice by engaging UX Designers as important creative professionals.UXD must align user needs with business values rather than functioning as purely artistic work.
- UXD focus: UXD remains an iterative, user-centered practice that depends on creativity, collaboration, problem solving, sense-making, and empathy.Designers translate user requirements into functionality and aesthetics through repeated iterations.
- UXD focus: GenAI can automate repetitive UX tasks such as wireframes, layout variations, and prototypes, potentially speeding iteration and broadening design exploration.The paper also identifies threats involving job displacement, ethics, lack of human touch, data quality, bias, and intellectual property.
2 RESEARCH QUESTIONS AND CONTRIBUTIONS
The paper asks how UX Designers perceive GenAI, how it may fit into current workflows, and what opportunities and risks they foresee for human–AI collaboration. Through interviews with 20 diverse UX Designers, it characterizes practice and finds assistive potential alongside serious concerns for junior designers, creativity, ownership, and access.
- Contributions: The paper contributes an overview of UXD practices, tools, limitations, workflows, and challenges that can serve as a baseline for assessing GenAI’s impact and utility.The authors also seek to understand how GenAI might fit or alter different workflows.
- Research questions: The study asks how UX Designers perceive GenAI, whether they can incorporate it into current workflows, and what future collaboration opportunities and risks they envision.These questions address both present practice and future human–AI collaboration.
- Approach: The researchers conducted in-depth interviews with 20 UX Designers spanning diverse experience levels and organizations from startups to enterprises with over 10,000 employees.Participants were mainly located across Europe and the United States.
- Findings: Experienced UX Designers are confident in their originality, creativity, and empathy, and view GenAI as assistive for repetitive and general tasks and productivity.The authors qualify this assistive role as applying to designers who are already skilled.
- Findings: Participants identify enjoyment and agency as uniquely human factors, with humans remaining final arbiters of AI alignment despite anticipated emergent abilities.The finding positions human judgment over AI output as central to collaboration.
- Risks and implications: Participants are especially concerned that junior designers may experience skill degradation and unemployment, lose systematic design-learning opportunities, and become trained primarily as prompters.They also raise copyright, ownership, creativity-exhaustion, homogenization, AI-literacy, and equal-access concerns.
- Implications: The authors argue that GenAI’s impact on UXD requires immediate adaptation and discuss responsible human–AI collaboration.The discussion is framed through responsible AI development.
3 BACKGROUND AND RELATED WORK
Prior work frames GenAI as affecting employment, creative practice, AI literacy, and human–AI collaboration, while this paper focuses specifically on UXD. The reviewed literature highlights both creative support and unresolved challenges involving authorship, control, uncertainty, and responsible design.
- AI and work: Research on AI and work estimates substantial exposure of occupations and tasks to language models, while other studies report productivity gains from AI tools.The cited literature includes estimates that at least 10% of U.S. work tasks could be affected and approximately 19% of workers could see at least half their tasks impacted.
- Creative practice: Studies of creative AI describe rapid design-space exploration, fluid collaboration, co-creation, prompt-based expression, and possible transformations in creative work and employment.These studies also examine authorship and ownership.
- Risks and governance: Related research identifies concerns about authorship, ownership, plagiarism, and several social impacts, including bias, privacy, disparate performance, and moderation costs.These concerns broaden the implications of GenAI beyond individual creative workflows.
- AI literacy: AI-literacy research defines competencies and design considerations for recognizing, evaluating, understanding, and critically interacting with AI systems.Examples include attention to human roles, ethics, explainability, transparency, and low barriers to entry.
- Human–AI collaboration: Human–AI collaboration research examines initiative, intent, control, authorship, and roles ranging from tools and assistants to mediators.One cited study finds that UI-supported initiative and control influence perceived authorship in text writing.
- Human-centered AI: Human-centered AI scholarship emphasizes human control, empowerment, governance, trustworthy systems, and collaboration between AI and HCI research.The literature also identifies AI capability uncertainty and output complexity as distinctive design challenges.
- Positioning: This paper extends prior work by analyzing how emerging GenAI affects UXD practices, adoption, and practitioner perceptions in depth.Its focus is one occupation closely connected to the HCI community.
4 METHOD
The study used semi-structured, one-on-one interviews with 20 UX Designers from varied backgrounds, companies, and experience levels, including demonstrations of current GenAI tools. Researchers analyzed transcripts through a deductive/inductive hybrid thematic process that produced six major themes.
- Participants and interviews: The researchers recruited UX Designers with diverse educational backgrounds, industry experience, company sizes, and locations across Europe and North America.Forty-eight people responded to screening, and 20 participated.
- Participants and interviews: The 20 participants completed online one-on-one interviews lasting approximately 60 minutes between April and July 2023.The sample included designers from startups, large enterprises, and companies with more than 10,000 employees.
- Participants and interviews: Participants varied in experience, with 11 reporting more than five years, two reporting three to five years, and seven reporting fewer than three years as UX Designers.All 20 regularly used ChatGPT; other participants used GPT-4, Midjourney, Notion AI, Miro AI, or Bard.
- Interview procedure: Interview questions covered backgrounds, workflows, UX challenges, GenAI attitudes, ethical and ownership concerns, collaboration roles, and desired future functionality.The procedure included demonstrations of GenAI tools for ideation, research-data synthesis, and front-end development.
- Interview procedure: Video demonstrations were used as probes to elicit reflection on GenAI tools and their relationship to participants’ practices, not to test prior use of specific tools.The demonstrations covered four UX-related activities.
- Analysis: Five researchers independently applied deductive codes, compared topics in a workshop, and converted the data into 415 digital statement cards.They then used inductive coding and affinity diagramming to surface sub-themes and consolidate categories.
- Analysis: The hybrid thematic analysis reached consensus through meetings and workshops rather than statistical inter-rater reliability and produced six major themes.The themes covered current tool limitations, UX challenges, GenAI attitudes, human–AI collaboration, ethics and ownership, and GenAI’s future in UXD.
- Analysis: The themes further included future collaboration, AI literacy and accessibility, desired AI roles, and challenges in defining AI-generated outcome metrics.Figures summarized the themes, commonly used UX tools, and workflow activities that GenAI can or cannot complete.
5 FINDINGS
The study maps UX Design workflows, tools, and organizational differences to establish how GenAI might fit diverse working environments. UX practice spans requirements gathering through ideation, prototyping, testing, iteration, and development, with workflows varying by company size.
- Implications for GenAI: GenAI adoption must account for different operational environments because one tool configuration may not fit large enterprises and smaller companies equally.The workflow baseline is intended to clarify which aspects of GenAI may be embraced or resisted by UX designers.
- UX Design workflows: The study identifies eight frequently used tool categories and eight typical steps in UX Design practice.The workflow extends from gathering requirements from business stakeholders to realizing solutions with the development team.
- UX Design tools: Figma was mentioned by all participants for high-fidelity prototyping and handoff, while Miro supported brainstorming and research-data synthesis.
- Organizational differences: Larger enterprises often divide UX Designer, Product Manager, and UX Researcher roles, whereas smaller companies often combine these responsibilities.Designers in smaller companies may conduct user research and spend more time securing stakeholder support for UX.
- Organizational differences: Smaller-company workflows tend to be shorter because fewer people are involved and communication costs are lower.Resource limitations may also cause smaller companies to omit or simplify workflow steps, including extensive user evaluations.
5.2 Theme 1: Limitations of Current UX Design Tools
Participants described current UX tools as fragmented and insufficiently connected across design and development. They highlighted frictional handoff, limited research support, and weak assistance for advanced design work.
- Frictional design handoff: Design specifications, assets, and documentation do not transition seamlessly from design tools into formats developers can readily access and understand.Participants wanted stronger gateways between designers and developers, including direct feeds from Figma prototypes into coding workflows.
- Fragmented tools: UX tools are scattered across functions, making transitions between tools such as Miro for low-fidelity work and Figma for high-fidelity work difficult.
- Limited research support: Participants wanted design tools to support usability testing and generate research data within the design tool itself.
- Lack of advanced design support: Current tools lack advanced support for complex animations, cross-platform designs, graphic design, and intelligent design systems.Participants specifically requested timeline-level animation control and systems that recommend and automatically adjust suitable components.
5.3 Theme 2: Challenges in Current UX Design
UX Designers face persistent challenges in communication, understanding end-users, balancing stakeholder and user needs, generating novel ideas, and accessing domain knowledge and resources. The passages also identify areas where GenAI may assist while emphasizing limits to substitution and reliability.
- Communication and collaboration: UX Designers often spend over half their time communicating and collaborating, while GenAI does not currently support this human-to-human work.
- Stakeholder alignment: Designers must balance stakeholder and client expectations with user needs while securing management buy-in and coordinating implementation with developers.
- Understanding end-users: Understanding users’ needs, goals, pain points, and broader context remains central and difficult in UX Design.Synthetic users and conversations are unlikely to currently substitute for designers’ deep understanding of end-users.
- Originality and constraints: Generating novel designs requires balancing established references, stakeholder expectations, user needs, business requirements, originality, copyright, and ownership.
- Domain knowledge and resources: Limited domain expertise and resources constrain access to specialized knowledge, experts, and large-scale user research or usability testing.Participants cited specialized fields and startup resource constraints as examples.
- Potential GenAI support: Text-based GenAI tools such as ChatGPT may provide immediate access to domain-specific information and resources despite current hallucinations.
5.4 Theme 3: UX Designers’ Attitudes toward GenAI
Participants viewed GenAI as useful for general and early-stage tasks but limited by generic outputs, manual effort, and uncertain accuracy. They valued human empathy and enjoyment while warning that excessive reliance could weaken creativity and judgment.
- Perceived benefits: All participants agreed that GenAI can support general tasks such as design ideation, basic UX writing, and basic coding.The tools were described as providing a starting point rather than completing the design process independently.
- Human oversight: Relying too heavily on GenAI to design products without human input was viewed as risky because outputs may not align with user needs and goals.Participants emphasized that current output quality still requires significant human input and review.
- Human capabilities: Participants considered human empathy and enjoyment in designing to be abilities that GenAI does not replace.One participant feared becoming primarily an editor and losing the ability to fill a blank canvas and empathize with users.
- Practicality and efficiency: GenAI outputs were often generic and still required substantial manual work, including redrawing, regenerating images, and crafting accurate prompts.
- Accuracy and reliability: Participants questioned GenAI accuracy and reliability, requiring human validation of generated designs, text, images, code, and stakeholder insights.
5.5 Theme 4: Human-AI Collaboration
Participants described GenAI primarily as an assistive tool for repetitive work, while retaining human responsibility for creativity, empathy, collaboration, and final decisions. They emphasized that human designers remain responsible for aligning and validating AI outputs.
- All participants characterized GenAI as a helper or assistant rather than a replacement for UX designers.
- GenAI was viewed as useful for repetitive tasks, including summarizing documents, generating prototypes and design variations, and supporting responsive design.
- Participants said UX work requiring stakeholder communication, empathy, user research, and user-centered functionality cannot be replaced by GenAI.
- Human designers were described as ambassadors for user needs and the ultimate decision-makers and arbiters of AI alignment.
- Participants linked human creativity and originality to understanding design contexts, empathizing with users, and innovating beyond AI’s reliance on past data.
- Participants considered human verification necessary throughout design and development because AI-generated content is not necessarily supported by verified research data.
5.6 Theme 5: Ethical Concerns and Ownership
Participants raised ethical concerns about ownership, copyright, privacy, transparency, access, skill degradation, and unemployment. They especially worried that automation could narrow designers’ skills and reduce junior designers’ opportunities to learn through practice.
- Participants identified copyright, privacy, data bias, transparency, skill degradation, unemployment, social disparity, and unequal access as ethical concerns.
- Participants worried that GenAI could worsen design plagiarism and make original creators harder to trace through modified or reproduced outputs.
- Some participants argued that humans should retain ownership when they refine prompts, contribute creativity, and oversee the process.
- Others recommended labeling AI contributions and documenting the tools and sources used to support transparent ownership attribution.
- Participants linked automation of essential design tasks to skill degradation, complacency, unemployment, and fewer learning opportunities for junior designers.
- Participants also worried that specialization and loss of oversight could narrow designers’ skills and jeopardize creativity.
- Privacy concerns included unintended disclosure of confidential information and uncertainty about how personal data would be retained for model training.
- Participants called for unbiased training data and equal access because policy, geography, education, and company resources may limit access to GenAI.
5.7 Theme 6: The Future of GenAI in UX Design
Participants envisioned GenAI as a collaborative design partner that could automate routine work, support research and handoffs, and make tools more accessible. They also identified the need for AI literacy and reliable ways to evaluate output quality, diversity, and innovation.
- Participants envisioned a collaborative system involving users, designers, and GenAI to minimize biases and consider development feasibility early.
- Participants called for AI education to prepare both junior and senior designers to use GenAI throughout UX practice.
- Participants wanted more visual, intuitive, and universally accessible AI tools rather than systems relying only on text-based inputs.
- Participants hoped GenAI would automate responsive components, animations, 3D content, documentation, and smoother handoffs to development teams.
- They wanted GenAI to support research, synthetic users, accessibility guidance, qualitative-data analysis, tool transitions, and design-to-development translation.
- They sought metrics and guidelines for evaluating AI outputs without overlooking crucial nuances relevant to design innovation and users.
- Participants warned that shared GenAI tools could produce similar designs and leave humans responsible for selecting and validating numerous outputs.
6 DISCUSSION
UX designers view GenAI as an assistive force whose value depends on preserving human creativity, enjoyment, agency, and control. The discussion highlights opportunities for productivity and collaboration alongside risks involving junior designers, ownership, creativity exhaustion, and misalignment with human values.
- GenAI’s role in UXD: Experienced UX designers regard GenAI as assistive, while remaining cautious about its effects on user-centered design and creative skills.They are confident in their own abilities but question whether GenAI can understand user-centered design and preserve creative expertise.
- Human-AI collaboration: GenAI may improve productivity by handling mundane work, but designers want collaboration to preserve human creativity, decision-making, and control.The proposed balance assigns routine tasks to AI while retaining human leadership over complex, creative, empathic, and ethical work.
- Human creativity and agency: Enjoyment and agency are recurring human factors that shape whether UX designers accept and continue using GenAI tools.Designers connect meaningful work with satisfaction and expect humans to remain active contributors and arbiters of decisions rather than passive editors.
- Creativity exhaustion: GenAI’s speed can pressure human creators to compete, producing mental and emotional strain that may contribute to creativity exhaustion.The discussion recommends emphasizing critical thinking and emotional depth rather than competing with AI’s speed, alongside support systems and education.
- Responsible and participatory AI: Responsible UXD collaboration requires evaluation frameworks, AI literacy, participatory involvement, and traceable approaches to copyright, ownership, and accountability.Designers support attribution of AI contributions, human participation in shaping tools, and oversight that helps align systems with human intentions and design practices.
- Skill degradation and employment: Junior designers face particular risks because replacing mundane tasks may remove opportunities to develop the experience needed for senior roles.Experienced designers also worry that emerging AI capabilities could eventually make parts of their expertise obsolete.
7 CONCLUSION
This study examines how UX Designers perceive GenAI in current practice and future human-AI collaboration. Experienced designers view GenAI as assistive while remaining confident in human originality, creativity, and empathy, but junior designers face risks including skill degradation, job replacement, and creativity exhaustion.
- The study investigated UX Designers’ perceptions of GenAI and envisioned opportunities and risks for future human-AI collaboration in UX Design practice.The researchers conducted in-depth interviews with 20 UX Designers from diverse experience backgrounds.
- Experienced designers remained confident in their originality, creativity, and empathic skills while viewing GenAI as primarily assistive.
- Humans were described as remaining in the driver seat over AI output, with enjoyment and agency identified as uniquely human factors.
- Junior designers may experience setbacks through skill degradation, job replacement, and creativity exhaustion.
- The authors identify responsible human-AI collaboration implications concerning copyright and ownership, human creativity, and AI literacy and access.
- They call for human and creative safeguarding mechanisms to help UX Design professionals adapt to GenAI’s disruptive impact.