Identifying Ethical Issues in AI Partners in Human-AI Co-Creation

Abstract

Human-AI co-creativity involves humans and AI collaborating on a shared creative product as partners. In many existing co-creative systems, users communicate with the AI using buttons or sliders. However, typically, the AI in co-creative systems cannot communicate back to humans, limiting their potential to be perceived as partners. This paper starts with an overview of a comparative study with 38 participants to explore the impact of AI-to-human communication on user perception and engagement in co-creative systems and the results show improved collaborative experience and user engagement with the system incorporating AI-to-human communication. The results also demonstrate that users perceive co-creative AI as more reliable, personal and intelligent when it can communicate with the users. The results indicate a need to identify potential ethical issues from an engaging communicating co-creative AI. Later in the paper, we present some potential ethical issues in human-AI co-creation and propose to use participatory design fiction as the research methodology to investigate the ethical issues associated with a co-creative AI that communicates with users.CCS Concepts: • Human-centered computing → User studies.

LAVE: LLM-Powered Agent Assistance and Language Augmentation for Video Editing

“…However, while AI holds significant promise for enhancing a user's creative abilities by managing certain aspects of the creative workflow, it also brings forward challenges and concerns such as user agency and trusts [30], the authenticity of the creation [50], potential creative biases [46,49], and ownership and credit attribution [10,21]. Our work builds upon existing literature in human-AI co-creation [8,14,21,24,30,32,40,55,57,73] and further contributes by developing a new AI system for video editing and studying its impact. Through the lens of LAVE, we examined the dynamics of user interactions with an LLM-based agent and explored the opportunities and challenges inherent in the proposed editing paradigm.…”