Computer ScienceFrontiers in Psychology
The paradox of creativity in generative AI: high performance, human-like bias, and limited differential evaluation
J. Desdevises
Generative AI like ChatGPT-4o can produce a high volume of ideas but still mirrors human fixation and struggles to distinguish original from conventional responses. This 2025 experiment, conducted by Joy Desdevises (OCTO Technology, Accenture, Paris, France), compared the model to 47 human participants and aggregated data from eight prior egg-task studies, revealing strong fluency yet limited originality evaluation and underscoring the need for human oversight.
Related Publications
Explore these studies to deepen your understanding
Adjacent work that informs or extends this paper's methodology and findings.
Computer Science
The paradox of creativity in generative AI: high performance, human-like bias, and limited differential evaluation
J. Desdevises
Psychology
Distance and grid-like codes support the navigation of abstract social space in the human brain
Z. Liang, S. Wu, et al.
Physics
High anisotropy in electrical and thermal conductivity through the design of aerogel-like superlattice (NaOH)<sub>0.5</sub>NbSe<sub>2</sub>
R. Sun, J. Deng, et al.
Medicine and Health
Differential routing and disposition of the long-chain saturated fatty acid palmitate in rodent vs human beta-cells
P. Thomas, C. Arden, et al.
