Melete: Exploring the Components of Mixed-Initiative Artificial Intelligence Pipelines for Level Design

Murturi, Sokol ORCID: https://orcid.org/0000-0001-9466-8981, Pellicone, Tony ORCID: https://orcid.org/0000-0002-9774-2953, Yee-King, Matthew ORCID: https://orcid.org/0000-0001-6606-2448 and Gillies, Marco ORCID: https://orcid.org/0000-0002-3100-9230 (2025) Melete: Exploring the Components of Mixed-Initiative Artificial Intelligence Pipelines for Level Design. In: Proceedings of the Fourth International Conference on Hybrid Human-Artificial Intelligence. Frontiers in Artificial Intelligence and Applications, 387unsure . IOS Press, Online. ISBN not known yet (In Press)

Preview

Text (CEUR_WS) CEUR_WS_Melete__Exploring_the_Components_of_Mixed_Initiative_Artificial_Intelligence_Pipelines_for_Level_Design (2).pdf - Accepted Version Available under License Creative Commons Attribution. Download (829kB) | Preview

Abstract / Summary

Recent advancements in artificial intelligence(AI) and human-computer interaction(HCI) have led to the creation
of innovative artifacts that bridge these fields, fostering creative collaboration between human authors and
artificial intelligence. These advancements have found applications across a wide range of academic disciplines.
Amongthesedevelopments, manytoolsandframeworkshaveemergedtosupportthedesignanddevelopment
of creative endeavors. However, there remains no clear consensus on how to structure these pipelines or what
components should be included.
To address this gap, we conducted a qualitative user study with twelve participants, examining how users
engage with key elements of a mixed-initiative artificial intelligence (MIAI) pipeline for game development,
including the procedural content generation (PCG) algorithm, its output, the user interface, playtesting, and the
overall pipeline. Through an inductive thematic analysis, we developed a mixed-initiative interaction model,
offering valuable insights into MIAI pipeline classification and guiding developers in designing more effective
and user-centred pipelines.

View Record (staff only)