Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
80 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
7 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

The Ink Splotch Effect: A Case Study on ChatGPT as a Co-Creative Game Designer (2403.02454v1)

Published 4 Mar 2024 in cs.AI

Abstract: This paper studies how LLMs can act as effective, high-level creative collaborators and ``muses'' for game design. We model the design of this study after the exercises artists use by looking at amorphous ink splotches for creative inspiration. Our goal is to determine whether AI-assistance can improve, hinder, or provide an alternative quality to games when compared to the creative intents implemented by human designers. The capabilities of LLMs as game designers are stress tested by placing it at the forefront of the decision making process. Three prototype games are designed across 3 different genres: (1) a minimalist base game, (2) a game with features and game feel elements added by a human game designer, and (3) a game with features and feel elements directly implemented from prompted outputs of the LLM, ChatGPT. A user study was conducted and participants were asked to blindly evaluate the quality and their preference of these games. We discuss both the development process of communicating creative intent to an AI chatbot and the synthesized open feedback of the participants. We use this data to determine both the benefits and shortcomings of AI in a more design-centric role.

ChatGPT as a Co-Creative Partner in Game Design: Evaluating the Ink Splotch Effect

The paper "The Ink Splotch Effect: A Case Study on ChatGPT as a Co-Creative Game Designer" investigates the role of LLMs, specifically ChatGPT, in the context of collaborative game design. The authors set out to explore whether AI can serve as an effective tool for creative inspiration and high-level collaboration akin to the artistic exercise of interpreting amorphous ink splotches. This experiment evaluates whether such AI can enhance, hinder, or offer a different alternative quality in game design compared to traditional human designers.

In this paper, the researchers focus on the potential of LLMs in a mixed-initiative creativity framework by using ChatGPT to assist with game design across three distinct genres: a space shooter, a platformer, and a roguelike game. Each genre underwent the creation of three versions: a baseline game, a human-enhanced game, and a game designed with the help of ChatGPT's suggestions. The authors then conducted a user paper where participants played and evaluated these games sensorially and qualitatively based on predefined metrics.

The results derived from participant feedback and rankings strongly demonstrated that human-designed games frequently outperformed AI-assisted ones in several metrics, including overall preference, game feel, and thematic cohesion. Human-designed games excelled in the categories requiring creativity and nuance, showcasing their ability to incorporate strong thematic processes and pragmatic reasoning. However, it is noteworthy that the ChatGPT-guided games received comparable ratings in the innovation category, suggesting that while AI might not yet mimic the depth of human creativity in game design, it can propose intriguing and diverse ideas.

In terms of methodology, the AI-driven games highlighted the capacity of ChatGPT to propose high-level feature suggestions that were pertinent and sometimes aligned with human creativity. However, ChatGPT experienced significant limitations in its contextual understanding and practical implementation within the Unity engine. For example, it struggled with in-depth game mechanisms such as procedural content generation and context-appropriate coding. This indicates that while ChatGPT can inspire creative directions and offer abstract guidance, reliance on it for detailed execution may require additional human expertise.

From a theoretical perspective, the implications of this research suggest a promising role for AI models like ChatGPT in augmenting the early stages of creative ideation for game developers. By serving in this role, AI can potentially operate as virtual muses, akin to the Bulletism art practice, offering a vast range of unclouded ideas that can prompt designers to think outside conventional frameworks. However, this paper emphasizes the continued importance of human oversight for developing thematic cohesion and pragmatic integrative processes.

Looking towards the future, the paper proposes expanding this research to other game engines with varying resources such as Godot or CryEngine and further iterating on real-life AI-human collaborative designs. This expansion will potentially provide deeper insights into how different contexts and limitations influence the dynamics of AI-driven creativity in fields beyond game design.

In conclusion, the research paper illustrates that while ChatGPT can enrich the creative process in game design through intriguing suggestions, it cannot replace the ingenuity and nuanced decision-making afforded by human designers. This interplay between human developers and AI holds potential for future explorations, as AI continues to develop as a capable, albeit complementary, tool in creative domains.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (23)
  1. Fostering creativity in the mixed-initiative evolutionary dungeon designer. In Proceedings of the 13th International Conference on the Foundations of Digital Games. 1–8.
  2. Team Artsapien. 2023. The art of bulletism: Understanding Salvador Dali’s vision. https://artsapien.com/2023/02/the-art-of-bulletism-understanding-salvador-dalis-vision/
  3. Lode Encoder: AI-constrained co-creativity. In 2021 IEEE Conference on Games (CoG). IEEE, 01–08.
  4. A Preliminary Study on a Conceptual Game Feature Generation and Recommendation System. arXiv preprint arXiv:2308.13538 (2023).
  5. The angelina videogame design system—part i. IEEE Transactions on Computational Intelligence and AI in Games 9, 2 (2016), 192–203.
  6. GDC. 2023. . YouTube. https://www.youtube.com/watch?v=fOqLuWml0UY
  7. Friend, collaborator, student, manager: How design of an ai-driven game level editor affects creators. In Proceedings of the 2019 CHI conference on human factors in computing systems. 1–13.
  8. Game mechanics telling stories? An experiment.. In FDG. Citeseer.
  9. Designer modeling for personalized game content creation tools. In Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, Vol. 9. 11–16.
  10. Computational game creativity. ICCC.
  11. Marian Mazzone and Ahmed Elgammal. 2019. Art, creativity, and the potential of artificial intelligence. In Arts, Vol. 8. MDPI, 26.
  12. The Five-Dollar Model: Generating Game Maps and Sprites from Sentence Embeddings. In Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, Vol. 19. 107–115.
  13. Mark J Nelson and Michael Mateas. 2008. An interactive game-design assistant. In Proceedings of the 13th international conference on Intelligent user interfaces. 90–98.
  14. Design-driven requirements for computationally co-creative game AI design tools. In Proceedings of the 16th International Conference on the Foundations of Digital Games. 1–12.
  15. Martin Pichlmair and Mads Johansen. 2021. Designing game feel: A survey. IEEE Transactions on Games 14, 2 (2021), 138–152.
  16. Prompt-Guided Level Generation. In Proceedings of the Companion Conference on Genetic and Evolutionary Computation. 179–182.
  17. Language as reality: a co-creative storytelling game experience in 1001 nights using generative AI. In Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, Vol. 19. 425–434.
  18. Steve Swink. 2008. Game feel: a game designer’s guide to virtual sensation. CRC press.
  19. Level Generation Through Large Language Models. In Proceedings of the 18th International Conference on the Foundations of Digital Games. 1–8.
  20. Julian Togelius and Jurgen Schmidhuber. 2008. An experiment in automatic game design. In 2008 IEEE Symposium On Computational Intelligence and Games. IEEE, 111–118.
  21. Game-o-matic: Generating videogames that represent ideas. In Proceedings of the The third workshop on Procedural Content Generation in Games. 1–8.
  22. Generating role-playing game quests with gpt language models. IEEE Transactions on Games (2022).
  23. ” An Adapt-or-Die Type of Situation”: Perception, Adoption, and Use of Text-To-Image-Generation AI by Game Industry Professionals. arXiv preprint arXiv:2302.12601 (2023).
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (5)
  1. Asad Anjum (1 paper)
  2. Yuting Li (21 papers)
  3. Noelle Law (2 papers)
  4. M Charity (13 papers)
  5. Julian Togelius (154 papers)
Citations (9)