Who" designs better? A competition among human, artificial intelligence and human-AI collaboration
Abstract
This research examines whether a machine, specifically artificial intelligence (AI), can be creative by comparing design solutions for a practical competition – a light fixture for a pediatric waiting room – among AI, collaboration efforts and a human designer. Amazon Mechanical Turk and Prolific workers observed the design solutions throughout the design process, from sketches (𝑆) to three-dimensional renderings (3𝐷) to fully developed models in virtual waiting rooms (𝑉𝑅). Using the well-established Creative Product Semantic Scale (CPSS), the workers rated each design solution in three distinctive stages – 𝑆, 3𝐷 and 𝑉𝑅 – on three criteria – novelty (freshness or newness), resolution (relevance and logic) and style (craftsmanship and desirability). Despite some demographic discrepancies, the workers expressed general senses of happiness and calmness, resonating with the competition’s requirements. Statistical results of CPSS ratings revealed that while AI excelled in style for 3𝐷, the human designer outperformed in novelty for both 𝑆 and 𝑉𝑅. Collaboration efforts surprisingly finished last. Such findings challenge current assumptions of AI’s creative ability in design research and highlight the need to be agile in the age of disruptive technologies. This research also offers guidance for product and interior designers and educators on thoughtfully integrating AI into the design process.
Study specs
- Authors
- KHT Vo
- Institution
- Indiana University
- Discipline
- Human-Computer Interaction
- Year
- 2025
- Human Data Platform
- Prolific
- Source
- View Source Google Scholar
Peer Review & Critical Discussion
Potential Selection Bias in 2023 Cohort
The participant pool shows a concerning overrepresentation of users from high-income demographics. Looking at Table 3, we can see that 78% of respondents had annual incomes above $75k, which significantly limits the generalizability of these findings to broader populations.
Non-naive Participants Issue
I've noticed a methodological concern regarding participant naivety. Given that Prolific users often complete multiple studies, there's a real risk that participants had prior exposure to similar experimental paradigms, which could confound the results.
RLHF Applicability to This Study Design
The implications for RLHF training pipelines are understated. If we accept the authors' conclusions about preference stability, this has direct consequences for how we should structure reward model training. The temporal decay effect described in Section 4.2 is particularly relevant.
Verify your expertise to join discussion
Create an account and verify your credentials to participate in peer discussions.