Friend, Collaborator, Student, Manager

Abstract

Machine learning advances have afforded an increase in algorithms capable of creating art, music, stories, games, and more. However, it is not yet well-understood how machine learning algorithms might best collaborate with people to support creative expression. To investigate how practicing designers perceive the role of AI in the creative process, we developed a game level design tool for Super Mario Bros.style games with a built-in AI level designer. In this paper we discuss our design of the Morai Maker intelligent tool through two mixed-methods studies with a total of over onehundred participants. Our findings are as follows: (1) level designers vary in their desired interactions with, and role of, the AI, (2) the AI prompted the level designers to alter their design practices, and (3) the level designers perceived the AI as having potential value in their design practice, varying based on their desired role for the AI.

Playful Probing: Towards Understanding the Interaction with Machine Learning in the Design of Maintenance Planning Tools

“…A better human-computer confluence can be achieved by enabling co-creation between the user and the artificial intelligence (AI) algorithm and putting explainability at the core of user autonomy and empowerment. Some studies [21,22] provide human-AI interaction guidelines, but explainable, accountable, and intelligible systems remain key challenges [23][24][25]. While progress has been made in explainability and interpretability [26][27][28], the design of AI interfaces focusing on co-creativity remains a challenge.…”

Image-to-Level: Generation and Repair

“…There have been some examples of PCGML co-creative systems. Guzdial et al (Guzdial et al 2019) employed a Deep Reinforcement Learning agent to collaborate with a user to produce a Super Mario Bros. level, acting as an equal partner. Schrum et al (Schrum et al 2020) employed a GAN trained on Super Mario Bros. and allowed users to explore the GAN's learned latent space.…”

“…To address the expertise problem, PCGML mixedinitiative systems have been created. These systems allow users to interact with and apply PCGML without ML expertise (Guzdial et al 2019;Schrum et al 2020). However, they still require a fairly high degree of effort to use effectively.…”

“…However, they still require a fairly high degree of effort to use effectively. For example, prior approaches require users to label existing data (Guzdial et al 2018), to learn the impacts of altering certain parameters (Schrum et al 2020) or to implicitly train an ML agent (Guzdial et al 2019). These requirements serve as barriers to more widespread adoption of mixed-initiative systems (Lai, Latham, and Leymarie 2020).…”

Designing Participatory AI: Creative Professionals’ Worries and Expectations about Generative AI

“…One step in the direction of participatory AI is to understand the needs of people and communities affected. Related research in this direction includes Singh et al, who explored which assumptions and expectations creative writers have for a supporting AI tool [24]; Guzdial et al, who explored designers' expectations for AI-driven gamelevel editors [12]; and Zhu et al, who argued for a better understanding of game designers' needs when co-creating with AI [29].…”