CoChat

Generative AI in Game Design: Enhancing Creativity or Constraining Innovation?

Open access
Alharthi, Sultan A. · 2025 · Journal of Intelligence

The DOI matches its Crossref record; 53 of the 53 references it cites were found in a registry or an index; 7 of the 61 sources its claims cite were checked for support.

Abstract

Generative AI tools are increasingly being integrated into game design and development workflows, offering new possibilities for creativity, efficiency, and innovation. This paper explores the evolving role of these tools from the perspective of game designers and developers, focusing on the benefits and challenges they present in fostering creativity. Through a mixed-method study, we conducted an online survey (n = 42) with game design professionals, followed by in-depth online interviews (n = 9), to investigate how generative AI influences the creative process, decision-making, and artistic vision. Our findings reveal that while generative AI accelerates ideation, enhances prototyping, and automates repetitive tasks, it also raises concerns about originality, creative dependency, and undermine of human-authored content. Future work will aim to address these challenges by investigating strategies to balance leveraging AI’s capabilities with preserving the integrity of human creativity. This includes developing collaborative human-AI workflows that maintain human oversight, designing systems that support rather than replace creative decision-making, and establishing ethical guidelines to ensure transparency, accountability, and authorship in AI-assisted content creation. By doing so, we aim to contribute to a more nuanced understanding of generative AI’s role in creative practices and its implications for the game design and development lifecycle.

Overview

CoChat Summary

TL;DR

This mixed-methods study examines how game designers and developers perceive generative AI in their workflows, using a survey (n = 42) and interviews (n = 9). It finds generative AI accelerates ideation, prototyping, and repetitive-task automation, while raising concerns about originality, authorship, creative dependency, and ethical/legal uncertainty.

Key findings

  • Ideation/brainstorming was the most common use (90.5%), followed by asset creation (83.3%), programming (73.8%), narrative/dialogue generation (66.7%), music/sound design (47.6%), procedural level generation (28.6%), and playtesting/QA (11.9%).
  • ChatGPT was the most used tool (90.5%), then Midjourney (76.2%), Copilot (69.0%), Gemini (64.3%), Scenario (61.9%), Artbreeder (52.4%), Stable Diffusion (45.2%), and Soundraw (35.7%).
  • Over 80% intended to use generative AI in future projects, and more than 75% agreed it helps them complete tasks more quickly, be more productive, and improve design quality.
  • Over 60% agreed generative AI might reduce originality, with a comparable proportion expressing ethical concerns; participants also worried about loss of the “human touch,” authorship, junior roles, and copyright/provenance.

Methods

  • Exploratory mixed-methods design: an online survey of game designers/developers (n = 42; age 19–45, M = 31, SD = 5.72; 66% male, 34% female), recruited from a WhatsApp game design and development community.
  • Follow-up semi-structured online interviews (n = 9; age 27–42, M = 36, SD = 4.98; 7 male, 2 female), recruited from survey participants.
  • Survey informed by TAM and UTAUT; data imported into ATLAS.ti and analyzed using exploratory thematic analysis (Braun and Clarke 2012), with inductive coding and analytic memos.

Limitations

  • Stated: small sample from the MENA region, recruited via a social media community, limiting statistical power/generalizability and risking selection bias; self-reported data may have response biases; coding was performed by a single researcher; no extensive comparison with traditional game design methods; not structured to compare organizational contexts.
  • Evident: the modest sample (42 survey participants, 9 interviewees) and single-coder thematic analysis mean findings are exploratory and not statistically generalizable, as the paper itself acknowledges.
  • Evident: reliance on self-report and perceptions rather than objective measures of AI’s impact, also acknowledged by the paper.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3390/jintelligence13060060

Cites

53 of 53 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 Participant demographics: region, education, role, and expertise level.
Figure 1 Participant demographics: region, education, role, and expertise level.CC BY · PubMed Central
Figure 2 Results from survey participants highlighting generative AI usage. (A) Specific tasks where generative AI was applied, such as ideation and brainstorming, asset creation, and narrative/dialogue generation. (B) Generative AI tools frequently utilized by participants, with ChatGPT, Stable Diffusion, and Midjourney emerging as the most popular choices.
Figure 2 Results from survey participants highlighting generative AI usage. (A) Specific tasks where generative AI was applied, such as ideation and brainstorming, asset creation, and narrative/dialogue generation. (B) Generative AI tools frequently utilized by participants, with ChatGPT, Stable Diffusion, and Midjourney emerging as the most popular choices.CC BY · PubMed Central
Figure 3 Example of game assets created by participants using generative AI tools.
Figure 3 Example of game assets created by participants using generative AI tools.CC BY · PubMed Central
Figure 4 Horizontal stacked bar chart representing the combined results of TAM and UTAUT. The chart visualizes the distribution of responses across the various questions, with color-coded segments indicating the percentage of participants selecting each level of agreement.
Figure 4 Horizontal stacked bar chart representing the combined results of TAM and UTAUT. The chart visualizes the distribution of responses across the various questions, with color-coded segments indicating the percentage of participants selecting each level of agreement.CC BY · PubMed Central

Verification

Claims the paper makes

5 Valid · 1 Not valid · 37 Not checked · 177 Own results and interpretations

95 more claims

Create a free account to see all 220 claims and to verify the 1 reference still open.

See all claims — free account
  • Probably not a substantial claim. 84
  • It may state two separate claims as one. 63
  • The quote only partly states this claim. 1

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Citations

Cites

Show all 53 references

Research works printed but not linked: 8

  1. Alharthi, Sultan A., Ben, Lafreniere, Tovi, Grossman, and George, Fitzmaurice (2025, May 10). TwoTorials: A Remote Cooperative Tutorial System for 3D Design Software In Graphics Interface 2022. Available online: https://openreview.net/forum?id=H2GICxFVaGc.
  2. Björk, Staffan, and Holopainen, Jussi (2005). Games and design patterns. The Game Design Reader: A Rules of Play Anthology, MIT Press.
  3. Choi, 2024, Investigating learner perceptions for effective teaching of generative AI-from a game development perspective, Journal of The Korea Society of Computer and Information, 29, 137
  4. Davis, 1989, Technology acceptance model, Journal of Management and Science, 35, 982
  5. Ersoy, Berkin (2024). AI AT PLAY: Exploratory Study on the Drivers Behind Generative AI Adoption for Game Art Asset Production. [Master’s thesis, Ghent University].
  6. Ratican, 2024, Adaptive worlds: Generative AI in game design and future of gaming, and interactive media, ISRG Journal of Arts, Humanities and Social Sciences, 2, 686
  7. Smith, Gillian, Whitehead, Jim, and Mateas, Michael (, January June). Tanagra: A mixed-initiative level design tool. Paper presented at the Fifth International Conference on the Foundations of Digital Games, Monterey, CA, USA.
  8. Wang, Qingyang (2024). Generative AI in Game Pedagogy. [Ph.D. thesis, Worcester Polytechnic Institute].

Cited by

See all 49

Metadata

Authors
Alharthi, S. A.
Affiliations
University of Jeddah (SA)
Corresponding author
Alharthi, Sultan A.
Venue
Journal of Intelligence
Publisher
MDPI AG
Published
2025-05-24
Volume
13
Issue
6
Pages
60
ISSN
2079-3200 (ISSN-L)
OpenAlex
W4410722094
Cited by
49
References
51
Citations by year
2026: 39 · 2025: 10
FWCI
12.61
Type
journal article
Language
en
Open access
Locations
  • Journal of Intelligence (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence in Games
  • Physical Sciences › Computer Science › Computer Graphics and Computer-Aided Design › Digital Media and Visual Art
  • Social Sciences › Sociology and Political Science › Digital Games and Media
Keywords
generative AI, game design, creativity, creative process, idea generation, prototype development, originality, human-AI collaboration, authorship, game development workflow

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-09-30, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-09-30 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.