Human-Centered AI Communication in Co-Creativity: An Initial Framework and Insights

Jeba Rezwana, Corey Ford · 2025

Read Human-Centered AI Communication in Co-Creativity: An Initial Framework and Insights by Jeba Rezwana, Corey Ford (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-18385.txt.

Introduction

Jeba Rezwana and Corey Ford sketch a frame for how smart tools should talk with people as they make things side by side, in a 2025 paper. Talk is the glue of joint work, yet many such tools talk poorly, which caps what the pair can do.

The frame, named FAICO, grew from a sweep of more than a hundred papers on joint making with smart tools. It names the key sides of machine talk and how each shapes the user's feel for the work. Two group talks with skilled users then tested the frame. People there backed a loop of give and take over one-way talk, and they stressed shared context. The authors turn the notes into steps for use and leads for more work.

Catalog record

CREATORJeba Rezwana, Corey Ford
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE64 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-18385

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Human-Centered AI Communication in Co-Creativity: An Initial Framework and Insights Jeba Rezwana, Corey Ford Abstract . Effective communication between AI and humans is essential for successful human-AI co-creation. However, many current co-creative AI systems lack effective communication, which limits their potential for collaboration. This paper presents the initial design of the Framework for AI Communication (FAICO) for co-creative AI, developed through a systematic review of 107 full-length papers. FAICO presents key aspects of AI communication and their impact on user experience, offering preliminary guidelines for designing human-centered AI communication. To improve the framework, we conducted a preliminary study with two focus groups involving skilled individuals in AI, HCI, and design. These sessions sought to understand participants’ preferences for AI communication, gather their perceptions of the framework, collect feedback for refinement, and explore its use in co-creative domains like collaborative writing and design. Our findings reveal a preference for a human-AI feedback loop over linear communication and emphasize the importance of context in fostering mutual understanding. Based on these insights, we propose actionable strategies for applying FAICO in practice and future directions, marking the first step toward developing comprehensive guidelines for designing effective human-centered AI communication in co-creation. Conference: Creativity and Cognition; June 23–25, 2025; Virtual, United Kingdom Creativity and Cognition (C&C ’25), June 23–25, 2025, Virtual, United Kingdom DOI: 10.1145/3698061.3726932 ISBN: 979-8-4007-1289-0/2025/06 CCS: Human-centered computing User centered design CCS: Human-centered computing Empirical studies in HCI 1. Introduction Human-AI co-creativity involves collaborative efforts between humans and AI in creative processes to produce artifacts, ideas, or performances ( Davis, 2013 ) , with research suggesting that the creativity resulting from this co-creation often surpasses the original intentions and creativity of the individual contributors ( Liapis et al., 2014 ) . As generative AI (GenAI) systems with co-creative abilities, such as ChatGPT ( ope, linea ) , DALL-E ( ope, lineb ) , Runway ( RunwayML, 2024 ) , and Midjourney ( mid, line ) , become increasingly popular, human-AI co-creativity is gaining traction in mainstream applications. As humans closely interact with AI in co-creativity, interaction has been suggested as an essential aspect of effective co-creative AI in the literature ( Bown, 2015 ; Kantosalo et al., 2020 ) . As a result, the next frontier of co-creative AI should have good collaborative skills in addition to algorithmic competence ( Rezwana and Maher, 2022a ) . However, designing effective co-creative AI presents several challenges due to the dynamic nature of human-AI interaction ( Davis et al., 2016 ; Kantosalo et al., 2014 ) . For example, co-creative AI must be capable of adapting to diverse and spontaneous interaction styles, allowing co-creative products to evolve dynamically across various stages of the creative process ( Rezwana and Maher, 2022a ) . Communication is an essential component in human-AI co-creativity for the co-regulation of the collaborators ( Bown et al., 2020 ) . In this paper, we define AI Communication as where an AI interacts with humans purposefully and directly to convey information. We distinguish AI Communication as a component of the interaction between collaborators (humans and AI) rather than interaction with the shared creative product itself. This distinction aligns with the framework proposed by Rezwana and Maher (2022a) in the co-creativity literature, where AI Communication is separate from AI contribution types and creative processes that influence the final output. Our focus is specifically on direct communication, excluding indirect communication through sensemaking that occurs through creative contributions. For instance, in a co-creative drawing scenario, AI might provide feedback on users’ contributions or proactively suggest design improvements, such as saying, “The house you drew is a little out of proportion. Would you like to fix it, or should I adjust it for you?” The user can choose to incorporate or disregard this suggestion, which distinguishes AI communication from AI’s contribution to the creative process, such as drawing a tree next to the user’s house. Our concept of AI communication aligns with the principles of explainable AI ( Kozierok et al., 2021 ; Gunning, 2016 ) , especially in how AI communication can offer explanations within co-creative AI contexts ( Bryan-Kinns et al., 2023 ) . To enable effective co-creation, Mamykina et al. (2002) emphasizes that co-creative AI should offer feedback and critique on user contributions in a manner similar to human collaborators in a team setting. Research in HCI has shown that when technology directly communicates with people, it is often perceived as an independent ‘social actor’ ( Nass et al., 1994 ) , which enhances user perception of AI ( Rezwana and Maher, 2022b ) . As such, AI Communication plays a crucial role in enabling co-creative AI to be viewed as an equal partner in the creative process ( McMillan and Hwang, 2002 ; Rezwana and Maher, 2021 ) , thereby improving user experience. Bidirectional exchange of ideas and information between users and AI can create moments of serendipity and novelty ( Lawton et al., 2023 ) , which are essential components of creativity ( Grace et al., 2015 ; Rezwana and Maher, 2024 ) . However, designing AI Communication remains a complex challenge due to the dynamic and evolving nature of human-AI collaboration ( Kantosalo et al., 2014 ) , which involves managing nuanced interactions and accommodating diverse user needs ( Weld and Bansal, 2018 ) . This highlights a research gap for tools, such as frameworks, to facilitate effective AI Communication in human-AI co-creativity. Motivated by the research gap, in this paper, we introduce a Framework for AI Communication (FAICO) in co-creativity to improve user experience as the first step towards designing effective AI Communication in co-creative contexts. Through a systematic review of interdisciplinary literature across HCI (107 full-length papers from the ACM Digital Library), we identified key aspects of AI Communication that we suggest should be considered in co-creativity and their influence on user experience, forming FAICO. We then present the findings of a preliminary user study in the form of two focus groups to gather feedback on FAICO in designing AI Communication for co-creative contexts. Our main contributions are: • A Framework for AI Communication (FAICO) for designing and interpreting AI communication . FAICO identifies key aspects of AI Communication that could be considered for effective co-creation and their influence on user experience. • Findings from a formative user study with participants skilled in AI, HCI and design, providing feedback on the framework and insights on user preferences for AI Communication. • Presentation of use-cases of our framework into practical tools for the community and future research directions. The paper is organized as follows. We first give background on computational creativity research, human-AI interaction processes and communication in human-AI co-creativity. We then describe our literature review methodology and introduce our framework. Next, we present the user study in Section 5. We close by discussing our findings, suggesting improvements and presenting use cases of the framework. 2. Background Co-creative systems originated from the concept of combining standalone generative AI (GenAI) systems with creativity support tools to create a system where both humans and AI take initiative in the creative process and interact as co-creators ( Kantosalo and Jordanous, 2020 ) . The terms human-AI co-creation and human-AI co-creativity are often used interchangeably

Read the complete text