Developments in artificial intelligence are currently dominated by the deep learning technology, which generates deep neural networks, trained on large data sets, which excel at pattern recognition. Variants of the “classic” deep neural networks have the ability to generate new data with statistical properties similar to the training set. Generative Adversarial Networks (GANs), such as used by DALL-E and Midjourney, may be used to generate original visual artworks based on a textual description of the desired output. Autoregressive language models, such as used by ChatGPT, use deep learning to produce text that is often indistinguishable from human-created text. Moreover, artificial intelligence techniques have been used to successfully generate music for many years, and researchers have also experimented with using deep learning to create cooking recipes, personalized fragrances, fashion, and more.

Despite the sometimes astonishing products of such creative artificial intelligence, the results are usually lacking in meaning. While DALL-E and Midjourney produce images that seem impressive, upon further inspection they contain many mistakes which humans would avoid. While ChatGPT can generate human-sounding text in a conversation, it often produces utter nonsense, and cannot write an original coherent story. And, as our own explorations of such techniques during Dagstuhl Seminar 22251 showed, GANs may produce computer game content which looks reasonable at first glance, but is ultimately neither functional nor playable.

While the product of creative artificial intelligence can often be used as a strong basis for humans to build upon, and may as such speed up the creative process, human intelligence and human creativity are almost always a necessary ingredient of the creative process. Moreover, the more relevant the meaning, purpose, and functionality of the product are, the less the creative process benefits from the involvement of artificial intelligence.

Game design and implementation are tasks which require a high amount of creativity, and which must lead to products which require a high amount of fine-tuned functionality. For example, a game “level” should not only look appealing, it should also be playable and it should be interesting to play. These are not features which can be acquired simply by “training on big data”, which is what most developments in modern artificial intelligence are based on.

The goal of this Dagstuhl Seminar is to investigate to what extent modern artificial intelligence techniques can produce meaningful and functional game content, and what changes to or extensions of these techniques can improve this AI-driven creative process. This seminar will bring together scientists, researchers, and industry professionals (both junior and senior) who specialize in Artificial Intelligence, Computational Creativity, Procedural Content Generation, and Game Development. The seminar will mostly focus on working groups, where small groups of participants work on a particular topic for half-a-day or a full day. This exploration may consist of a discussion or of the building of a prototype, but should always produce some results. The collaboration between scientists and industry professionals should lead to products which progress the state-of-the-art in computational creativity.

Creative Commons BY 4.0

Duygu Cakmak, Setareh Maghsudi, Diego Perez Liebana, and Pieter Spronck