De-skilling, Cognitive Offloading, and Misplaced Responsibilities: Potential Ironies of AI-Assisted Design Conference: Extended Abstracts of the CHI Conference on Human Factors in Computing Systems; April 26-May 1, 2025; Yokohama, Japan Extended Abstracts of the CHI Conference on Human Factors in Co

Prakash Shukla, Phuong Ngo Ngoc Bui, Sean Levy, Maxwell Kowalski, Ali Baigelenov, Paul Parsons · 2025

Read De-skilling, Cognitive Offloading, and Misplaced Responsibilities: Potential Ironies of AI-Assisted Design Conference: Extended Abstracts of the CHI Conference on Human Factors in Computing Systems; April 26-May 1, 2025; Yokohama, Japan Extended Abstracts of the CHI Conference on Human Factors in Co by Prakash Shukla, Phuong Ngo Ngoc Bui, Sean Levy, Maxwell Kowalski, Ali Baigelenov, Paul Parsons (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-03924.txt.

Introduction

Prakash Shukla and co-authors warn that design help from smart tools can dull the very skills it aids, in a study from 2025. Such tools now draft screens and text at a tap. Gains in speed are real, but old lore on auto aids lists ironies: skills fade, thought is offloaded, blame lands wrong.

The team read more than 120 posts and chats from design forums. Hands-on designers told of both hope and dread: joy at less grind and more spark, fear of lean on the tool, thin thought, and lost craft. The team ties those tales to the old irony lore and to the split of work between human and tool. Their charge: judge the tool by craft years from now, not by this week's speed.

Catalog record

CREATORPrakash Shukla, Phuong Ngo Ngoc Bui, Sean Levy, Maxwell Kowalski, Ali Baigelenov, Paul Parsons
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE36 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-03924

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

De-skilling, Cognitive Offloading, and Misplaced Responsibilities: Potential Ironies of AI-Assisted Design Conference: Extended Abstracts of the CHI Conference on Human Factors in Computing Systems; April 26-May 1, 2025; Yokohama, Japan Extended Abstracts of the CHI Conference on Human Factors in Computing Systems (CHI EA ’25), April 26-May 1, 2025, Yokohama, Japan DOI: 10.1145/3706599.3719931 ISBN: 979-8-4007-1395-8/2025/04 CCS: Human-centered computing Empirical studies in HCI CCS: Human-centered computing HCI theory, concepts and models Prakash Shukla, Phuong Ngo Ngoc Bui, Sean Levy, Maxwell Kowalski, Ali Baigelenov, Paul Parsons Abstract . The rapid adoption of generative AI (GenAI) in design has sparked discussions about its benefits and unintended consequences. While AI is often framed as a tool for enhancing productivity by automating routine tasks, historical research on automation warns of paradoxical effects, such as de-skilling and misplaced responsibilities. To assess UX practitioners’ perceptions of AI, we analyzed over 120 articles and discussions from UX-focused subreddits. Our findings indicate that while practitioners express optimism about AI reducing repetitive work and augmenting creativity, they also highlight concerns about over-reliance, cognitive offloading, and the erosion of critical design skills. Drawing from human-automation interaction literature, we discuss how these perspectives align with well-documented automation ironies and function allocation challenges. We argue that UX professionals should critically evaluate AI’s role beyond immediate productivity gains and consider its long-term implications for creative autonomy and expertise. This study contributes empirical insights into practitioners’ perspectives and links them to broader debates on automation in design. 1. Introduction The rapid advancement of generative AI (GenAI) has led to significant developments across various creative domains. Large Language Models (LLMs) can now engage in sophisticated conversations, generate text, produce high-quality images and videos, and even create interactive UI mockups for websites and applications. Many widely used UX design tools, such as Figma, Miro, and Framer, are integrating AI into their platforms, further accelerating discussions about the role and impact of AI in design. In the UX design community, practitioners are actively discussing how AI-driven tools will reshape their roles and workflows. Concurrently, researchers are examining the implications of AI for design practices, focusing on both the opportunities and challenges posed by AI integration ( Stige et al. 2023 ; Inie et al. 2023 ; Li et al. 2024 ; Zimmerman 2009 ; Kalving et al. 2024 ; Khan et al. 2025 ; Olsson and Väänänen 2021 ) . While AI has the potential to automate routine tasks, most creative professionals recognize that human input remains essential to the design process ( Inie et al. 2023 ) . Despite the promise of AI-driven tools in design, their integration into cognitive work introduces complexities that extend beyond simple efficiency gains. Predicting how emerging technologies will reshape professional practice is inherently difficult, as these changes often bring unintended consequences affecting work environments, responsibilities, coordination, and interactions with tools ( Dekker et al. 2002 ) . History has shown that automation can lead to unexpected challenges, sometimes creating inefficiencies rather than resolving them. Bainbridge ( Bainbridge 1983 ) famously outlined several ironies of automation in process industries, many of which remain relevant to contemporary AI integration ( Endsley 2023 ) . With prior cycles of automation hype and setbacks in mind, we aim to investigate whether UX designers are aware of the potential ironies associated with LLM-driven AI. Recent studies have begun examining the paradoxical effects of GenAI on practitioners, emphasizing how automation may, in some cases, hinder productivity rather than enhance it. For example, Simkute et al. ( Simkute et al. 2024 ) draw on human factors research to analyze the usability challenges of GenAI and propose strategies to mitigate its potential drawbacks. However, there remains an opportunity to build upon decades of research in human-automation interaction to better understand its implications for UX design practitioners. In this work, we explore the following questions: What are UX designers’ perspectives on GenAI’s impact on UX practice? What potential ironies and concerns may arise as AI becomes integrated into the design process? To address these questions, we analyze data from two sources: blog posts written by UX practitioners and discussions on UX-related subreddits. Blogs have become a widely used medium for practitioners to share insights and opinions in an accessible and cost-effective manner ( Herring et al. 2005 ) . Platforms such as Medium, Smashing Magazine, and personal blogs host an expanding community of UX professionals actively contributing to design discourse. Additionally, large corporations engage in these discussions through company web pages, publishing articles authored by design team members and leaders ( noa 2024b ) . Prior research has demonstrated the value of subreddit analysis in capturing UX practitioners’ perspectives ( Shukla et al. 2024 ; Kou et al. 2018 ) . By examining both blogs and subreddit discussions, we gain a broad and diverse view of current attitudes toward AI in UX design. Furthermore, we draw on literature from the history of automation to examine potential ironies and unintended consequences of AI adoption in UX practice. This historical perspective helps contextualize the challenges and complexities of integrating AI into creative and cognitive work. 2. Background 2.1. Integrating GenAI into Design Practices The integration of GenAI into design practice offers new opportunities while simultaneously raising important questions about the nature of design work. The possible effects have sparked both enthusiasm and apprehension among practitioners ( Kalving et al. 2024 ; Li et al. 2024 ) . Some discussions have revolved around the potential of GenAI to augment human capabilities and automate mundane tasks ( Kalving et al. 2024 ; Khan et al. 2025 ; Olsson and Väänänen 2021 ) . However, this integration introduces ethical dilemmas, practical challenges, and questions about the evolving roles of designers ( Chaudhry 2024 ) . One central discussion revolves around AI as a collaborative partner versus a mere tool. While some designers view AI as a soundboard or co-worker, others insist on maintaining human control over the creative process ( Kalving et al. 2024 ; Khan et al. 2025 ) . Concerns exist regarding the potential for AI to diminish the designer’s role and erode essential skills, particularly among junior designers ( Li et al. 2024 ) . Ethical considerations are significantly important, with discussions focusing on copyright issues, the legitimacy of AI-generated content, and the perpetuation of biases ( Chaudhry 2024 ; Kalving et al. 2024 ; Knearem et al. 2023 ) . The ‘black-box’ nature of AI introduces variability in control, making it challenging for professionals to predict outcomes accurately ( Law and Varanasi 2025 ) , raising ethical questions about who maintains responsibility for outputs when processes are not transparent. A key tension lies in balancing efficiency with creative freedom. While AI can streamline processes such as data collection and prototyping, there is an ongoing debate about whether AI intervention constrains creative exploration ( Khan et al. 2025 ) . For instance, AI-enabled design tools are typically focused on graphical outputs and fail to provide support for other aspects of creativity like those relating to design thinking ( Khan et al. 2025 ; Uusitalo et al. 2024 ) . Additionally, because AI tools can quickly generate high-fidelity prototypes, processes relating to creativity that occur early in the design process, and typically with

Read the complete text