Generative Artificial Intelligence (GenAI) is currently reshaping game development practices, production pipelines, and value networks in an unprecedentedly pervasive manner with cascading consequences remaining unclear. In the last five years since GenAI’s inception, a growing body of qualitative research has explored these early transformations from different settings and demographic angles. However, these studies often contextualise and consolidate their findings weakly with related work; for research to keep up with and support stakeholders in this development, the current moment calls for a synthesis of the findings emerged thus far. Here, we address this need through a qualitative research synthesis via meta-ethnography. We followed PRISMA-S to systematically search the relevant literature from 2020-2025, including major HCI and games research databases. We then synthesised the ten eligible studies, conducting reciprocal translation and line-of-argument synthesis guided by eMERGe, informed by CASP quality appraisal. We identified nine overarching themes, provide recommendations, and contextualise our insights in wider game production trajectories. With this work, we seek to provide practitioners, researchers and policy-makers with grounded insights to guide practice, research and governance.

Generative AI tools for image creation are now mainstream, yet we know little about when and why observers judge them as “creative”. Previous human-robot interaction research suggests that revealing the creation process can raise perceived machine creativity and points that observer differences may moderate this effect. We take these observations from physical robots to the bigger domain of virtual text-to-image diffusion systems by manipulating perceptual evidence (PE), i.e., interface-visible cues about the generation process. We report two preregistered online experiments looking into PE and observer individual differences. Study 1 (N=298) used a within-subjects manipulation comparing Product (final image only) to Product+Process (adding a short animation of the denoising process). Study 2 (N=295) added a between-subjects tutorial (diffusion vs. control) in a 2 × 2 mixed design. The tutorial briefly explained how diffusion models generate images, intended to raise system-specific literacy. Contrary to previous work, confirmatory analyses found no average effect of showing Process on creativity, and no tutorial effect. Exploratory analyses revealed that general AI literacy moderated the PE contrast, i.e., at lower literacy, observing process tended to lower creativity ratings; at higher literacy, it tended to raise them. Moreover, attitudes toward AI and art interest were positively associated with creativity ratings. Thematic analysis of open-ended responses indicated potential reasons for the lack of overall PE effect. Taken together, these converging quantitative and qualitative findings indicate that individual differences systematically shape creativity judgments of text-to-image GenAI and, in our setting, exert stronger and more reliable influence than PE alone. For design, this implies that process visualizations could help some audiences more than others. Interfaces that adapt to literacy and attitudes, or that pair process views with contextual explanation calibrated to user background, could be more likely to shift judgments than one-size-fits-all depictions of generation.

While creative artificial intelligence (AI) is becoming integral to our lives, we know little about what makes us call AI “creative”. Informed by prior theoretical and empirical work, we investigate how perceiving evidence of a creative act beyond the final product affects our assessment of robot creativity. We study embodiment morphology as a potential moderator of this relationship, informing a 3 × 2 factorial design. In two lab experiments on visual art, participants (N = 30 + 60) assessed drawings produced by two physical robots with different morphologies, under exposure to product, process and producer as three levels of perceptual evidence. The data supports that the human assessment of robot creativity is significantly higher the more is revealed beyond the product about the creation process, and eventually the producer. We find no significant effects of embodiment morphology, contrasting existing hypotheses and offering a more detailed understanding for future work. The latter is also informed by additional exploratory analyses revealing factors potentially influencing creativity assessments, including perceived robot likeability and participants’ experience with robotics and AI. Our insights empirically ground existing design patterns, foster fairness and validity in system comparisons, and contribute to a deeper understanding of our relationship with creative AI and thus its adoption in society.

Similarity estimation is essential for many game AI applications, from the procedural generation of distinct assets to automated exploration with game-playing agents. While similarity metrics often substitute human evaluation, their alignment with our judgement is unclear. Consequently, the result of their application can fail human expectations, leading to e.g. unappreciated content or unbelievable agent behaviour. We alleviate this gap through a multi-factorial study of two tile-based games in two representations, where participants (N=456) judged the similarity of level triplets. Based on this data, we construct domain-specific perceptual spaces, encoding similarity-relevant attributes. We compare 12 metrics to these spaces and evaluate their approximation quality through several quantitative lenses. Moreover, we conduct a qualitative labelling study to identify the features underlying the human similarity judgement in this popular genre. Our findings inform the selection of existing metrics and highlight requirements for the design of new similarity metrics benefiting game development and research.

Quests represent an integral part of role-playing games (RPGs). While evocative, narrative-rich quests are still mostly hand-authored, player demands towards more and richer game content, as well as business requirements for continuous player engagement necessitate alternative, procedural quest generation methods. While existing methods produce mostly uninteresting, mechanical quest descriptions, recent advances in AI have brought forth generative language models with promising computational storytelling capabilities. We leverage two of the most successful Transformer models, GPT-2 and GPT-3, to procedurally generate RPG video game quest descriptions. We gathered, processed and openly published a data set of 978 quests and their descriptions from six RPGs. We fine-tuned GPT-2 on this data set with a range of optimizations informed by several mini studies. We validated the resulting Quest-GPT-2 model via an online user study involving 349 RPG players. Our results indicate that one in five quest descriptions would be deemed acceptable by a human critic, yet the variation in quality across individual quests is large. We provide recommendations on current applications of Quest-GPT-2. This is complemented by case-studies on GPT-3 to highlight the future potential of state-of-the-art natural language models for quest generation.

Many human-AI authoring tools are used in a playful way, while being primarily designed for task-achievement—not playfulness. We argue that playfulness is an important yet overlooked factor of user behaviour and experience when interacting with such tools. Motivating and rewarding playfulness as an exploratory, task-agnostic, open, and subversive attitude can support the satisfaction of more diverse user goals, and have a strong, positive effect on the user experience, the emerging human-AI interaction, and the resulting artefact. In this paper, we motivate the importance of playfulness as user experience in human-AI authoring tools, and propose concrete strategies to design for playfulness in the human user through UI design, in the AI through algorithms, or through interventions to their dialog. We conclude with an outlook of the research agenda.

Text-to-image generation (TTIG) models, a recent addition to creative AI, can generate images based on a text description. These models have begun to rival the work of professional creatives, and sparked discussions on the future of creative work, loss of jobs, and copyright issues, amongst other important implications. To support the sustainable adoption of TTIG, we must provide rich, reliable and transparent insights into how professionals perceive, adopt and use TTIG. Crucially though, the public debate is shallow, narrow and lacking transparency, while academic work has focused on studying the use of TTIG in a general artist population, but not on the perceptions and attitudes of professionals in a specific industry. In this paper, we contribute a qualitative, exploratory interview study on TTIG in the Finnish videogame industry. Through a Template Analysis on semi-structured interviews with 14 game professionals, we reveal 12 overarching themes, structured into 39 sub-themes on professionals' perception, adoption and use of TTIG in games industry practice. Experiencing (yet another) change of roles and creative processes, our participants' reflections can inform discussions within the industry, be used by policymakers to inform urgently needed legislation, and support researchers in games, HCI and AI to support the sustainable, professional use of TTIG, and foster games as cultural artefacts.

The GDMC AI settlement generation challenge is a procedural content generation (PCG) competition about producing an algorithm that can create a settlement in the game Minecraft. In contrast to the majority of AI competitions, the GDMC entries are evaluated by human experts on several criteria such as adaptability, functionality, evocative narrative, and visual aesthetics – all of which represent challenges to state-of-the-art PCG systems. This paper contains a collection of written experiences with this competition, by participants, judges, organizers and advisors. We asked people to reflect both on the artifacts themselves, and on the competition in general. The aim of this paper is to offer a shareable and edited collection of experiences and qualitative feedback which have the potential to push forward PCG and computational creativity, but would be lost once the individual assessments are compressed to scalar ratings. We reflect upon organizational issues for AI competitions, and discuss the future of the GDMC competition.

The accurate and personalized estimation of task difficulty provides many opportunities for optimizing user experience. However, user diversity makes such difficulty estimation hard, in that empirical measurements from some user sample do not necessarily generalize to others.

Cutscenes form an integral part of many video games, but their creation is costly, time-consuming, and requires skills that many game developers lack. While AI has been leveraged to semi-automate cutscene production, the results typically lack the internal consistency and uniformity in style that is characteristic of professional human directors. We overcome this shortcoming with Cine-AI, an open-source procedural cinematography toolset capable of generating in-game cutscenes in the style of eminent human directors. Implemented in the popular game engine Unity, Cine-AI features a novel timeline and storyboard interface for design-time manipulation, combined with runtime cinematography automation. Via two user studies, each employing quantitative and qualitative measures, we demonstrate that Cine-AI generates cutscenes that people correctly associate with a target director, while providing above-average usability. Our director imitation dataset is publicly available, and can be extended by users and film enthusiasts.

This paper presents a novel approach to automated playtesting for the prediction of human player behavior and experience. We have previously demonstrated that Deep Reinforcement Learning (DRL) game-playing agents can predict both game difficulty and player engagement, operationalized as average pass and churn rates. We improve this approach by enhancing DRL with Monte Carlo Tree Search (MCTS). We also motivate an enhanced selection strategy for predictor features, based on the observation that an AI agent's best-case performance can yield stronger correlations with human data than the agent's average performance. Both additions consistently improve the prediction accuracy, and the DRL-enhanced MCTS outperforms both DRL and vanilla MCTS in the hardest levels. We conclude that player modelling via automated playtesting can benefit from combining DRL and MCTS. Moreover, it can be worthwhile to investigate a subset of repeated best AI agent runs, if AI gameplay does not yield good predictions on average.

This paper presents a review of intrinsic motivation in player modeling, with a focus on simulation-based game testing. Modern AI agents can learn to win many games; from a game testing perspective, a remaining research problem is how to model the aspects of human player behavior not explained by purely rational and goal-driven decision making. A major piece of this puzzle is constituted by intrinsic motivations, i.e., psychological needs that drive behavior without extrinsic reinforcement such as game score. We first review the common intrinsic motivations discussed in player psychology research and artificial intelligence, and then proceed to systematically review how the various motivations have been implemented in simulated player agents. Our work reveals that although motivations such as competence and curiosity have been studied in AI, work on utilizing them in simulation-based game testing is sparse, and other motivations such as social relatedness, immersion, and domination appear particularly underexplored.

Christian Guckelsberger and Christoph Salge, 'Does Empowerment Maximisation Allow for Enactive Artificial Agents?' in Proceedings of the Fifteenth International Conference on the Synthesis and Simulation of Living Systems (Alife 2016), Cancun, Mexico, 4-8 July 2016. Carlos Gershenson, Tom Froese, Jesus M. Siqueiros, Wendy Aguilar, Eduardo J. Izquierdo and Hiroki Sayama eds., ISBN 9780262339360. Published by MIT Press.

Social media represents a rich source of upto-date information about events such as incidents.The sheer amount of available information makes machine learning approaches a necessity for further processing.This learning problem is often concerned with regionally restricted datasets such as data from only one city.Because social media data such as tweets varies considerably across different cities, the training of efficient models requires labeling data from each city of interest, which is costly and time consuming.In this study, we investigate which features are most suitable for training generalizable models, i.e., models that show good performance across different datasets.We reimplemented the most popular features from the state of the art in addition to other novel approaches, and evaluated them on data from ten different cities.We show that many sophisticated features are not necessarily valuable for training a generalized model and are outperformed by classic features such as plain word-n-grams and character-n-grams.