Visual ideation mediated by generative artificial intelligence: Lexical diversity, prompt complexity, and generation settings in DiffusionDB
DOI:
https://doi.org/10.64747/2mgg3g14Keywords:
generative artificial intelligence, visual ideation, prompt engineering, computational creativity, content analysis, DiffusionDBAbstract
Prompt writing is an observable stage of ideation in text-to-image generative systems; however, its diversity, recurrence, and relationship with generation settings must be examined without equating textual complexity with psychological creativity. This study characterized visual ideation expressed in public DiffusionDB prompts through lexical diversity, structural elaboration, recurrence, and associations with generation parameters. We conducted a cross-sectional observational study using secondary public data. All 2,000,000 DiffusionDB 2M records were audited. Text analyses included 1,999,397 non-empty prompts, 1,522,692 unique normalized formulations, and a deterministic sample of 100,000 prompts for lexical metrics. Prompts had a median of 21 words and five segments. Repeated formulations accounted for 23.84% of records. The most frequent explicit markers were medium or technique (57.35%), style or artist (55.28%), and quality or detail (45.46%). The elaboration index had a median of two categories, and 60.65% of records combined at least two visual resources. Correlations between prompt length and generation settings were small within users (r = 0.12–0.16) and moderate across users (rho = 0.33–0.51). Prompts operated as comparatively rich compositional assemblages while relying on recurrent formulas. These findings describe early Stable Diffusion ideation practices; they do not measure human creativity or visual quality.
References
Covington, M. A., & McFall, J. D. (2010). Cutting the Gordian knot: The moving-average type-token ratio (MATTR). Journal of Quantitative Linguistics, 17(2), 94–100. https://doi.org/10.1080/09296171003643098
De Rosa Palmini, M.-T., & Cetinic, E. (2024). Patterns of creativity: How user input shapes AI-generated visual diversity [Preprint]. arXiv. https://arxiv.org/abs/2410.06768
Doshi, A. R., & Hauser, O. P. (2024). Generative AI enhances individual creativity but reduces the collective diversity of novel content. Science Advances, 10(28), eadn5290. https://doi.org/10.1126/sciadv.adn5290
Epstein, Z., Hertzmann, A., Akten, M., Farid, H., Fjeld, J., Frank, M. R., Groh, M., Herman, L., Leach, N., Mahari, R., Pentland, A. S., Russakovsky, O., Schroeder, H., & Smith, A. (2023). Art and the science of generative AI. Science, 380(6650), 1110–1111. https://doi.org/10.1126/science.adh4451
Feng, Y., Wang, X., Wong, K. K., Wang, S., Lu, Y., Zhu, M., Wang, B., & Chen, W. (2023). PromptMagician: Interactive prompt engineering for text-to-image creation. IEEE Transactions on Visualization and Computer Graphics, 1–11. https://doi.org/10.1109/TVCG.2023.3327168
Liu, V., & Chilton, L. B. (2022). Design guidelines for prompt engineering text-to-image generative models. In Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems (pp. 1–23). Association for Computing Machinery. https://doi.org/10.1145/3491102.3501825
McCarthy, P. M., & Jarvis, S. (2010). MTLD, vocd-D, and HD-D: A validation study of sophisticated approaches to lexical diversity assessment. Behavior Research Methods, 42(2), 381–392. https://doi.org/10.3758/BRM.42.2.381
Oppenlaender, J., Linder, R., & Silvennoinen, J. (2024). Prompting AI art: An investigation into the creative skill of prompt engineering. International Journal of Human–Computer Interaction. Advance online publication. https://doi.org/10.1080/10447318.2024.2431761
Ramesh, A., Dhariwal, P., Nichol, A., Chu, C., & Chen, M. (2022). Hierarchical text-conditional image generation with CLIP latents [Preprint]. arXiv. https://arxiv.org/abs/2204.06125
Rombach, R., Blattmann, A., Lorenz, D., Esser, P., & Ommer, B. (2022). High-resolution image synthesis with latent diffusion models. In 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 10674–10685). IEEE. https://doi.org/10.1109/CVPR52688.2022.01042
Saharia, C., Chan, W., Saxena, S., Li, L., Whang, J., Denton, E., Seyed Ghasemipour, S. K., Ayan, B. K., Mahdavi, S. S., Lopes, R. G., Salimans, T., Ho, J., Fleet, D. J., & Norouzi, M. (2022). Photorealistic text-to-image diffusion models with deep language understanding [Preprint]. arXiv. https://arxiv.org/abs/2205.11487
Wadinambiarachchi, S., Kelly, R. M., Pareek, S., Zhou, Q., & Velloso, E. (2024). The effects of generative AI on design fixation and divergent thinking. In Proceedings of the CHI Conference on Human Factors in Computing Systems (pp. 1–18). Association for Computing Machinery. https://doi.org/10.1145/3613904.3642919
Wang, Z. J., Montoya, E., Munechika, D., Yang, H., Hoover, B., & Chau, D. H. (2022). DiffusionDB: A large-scale prompt gallery dataset for text-to-image generative models [Preprint]. arXiv. https://arxiv.org/abs/2210.14896
Zhou, E., & Lee, D. (2024). Generative artificial intelligence, human creativity, and art. PNAS Nexus, 3(3), pgae052. https://doi.org/10.1093/pnasnexus/pgae052
Downloads
Published
Versions
- 2025-06-29 (2)
- 2025-06-29 (1)
Issue
Section
License
Copyright (c) 2025 MARIA FERNANDA RODRIGUEZ-RUIZ (Autor/a)
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors retain copyright and grant Sapiens Global the right of first publication. The article is distributed under CC BY 4.0, except where third-party material is identified under different terms.
