Integrating AI in Human-Human Collaborative Ideation

Technology has many roles to play in human cognition. It has been recently proposed that Cultural-Historical Activity Theory could be applied as an analytical lens to structure human-machine collaboration in creative contexts. In this position paper, we apply activity theory tenets to formalize possible co-creative scenarios and discuss the possible roles that AI can play in the collaborative process of the production of novel and useful artifacts. In addition to 'disembodied' artificial systems that emphasize creative ideation, we stress the role of physical presence and embodied joint action for studying how creative artifacts emerge from human and machine interactions with the social and material worlds. A future artificial agent that could be a partner in human creative work is envisioned as co-embodied and socially present.

[PDF]

This paper explores applications of augmented collective intelligence (ACI) beneficial to collaborative ideation. Design considerations are offered for an experiment that evaluates the performance of hybrid human–AI collectives. The experiment described brings humans and large language models (LLMs) together in order to ideate on increasingly complex topics. To facilitate ACI, a promising real-time collection tool called Polis is examined, including case studies from citizen engagement projects in Taiwan and Bowling Green, Kentucky. The authors go on to discuss three challenges to consider when designing an ACI experiment: topic selection, participant selection, and evaluation of results. AI systems may be better thought of as peers than as tools. The paper concludes that researchers should address the these challenges to conduct empirical studies of ACI in collaborative ideation.

[PDF]

In this position paper, we argue for the importance of exploring the role of empathy in interdisciplinary ideation, especially in the context of designing Human-AI co-creative systems. As part of our previous teaching and work, we have observed challenges in design teams especially because of designers’ interdisciplinarity. We reflect on the importance of fostering empathy in design thinking, not only towards end-users but also among designers themselves. AI systems have already demonstrated many benefits in terms of design thinking and ideation, however, the role of AI systems in relation to the key aspect of empathy in design thinking needs to be considered carefully.

[PDF]

In this position paper, we reflect on how large language models (LLMs) might support designers in ideation, concept generation, and assessment.We argue that the majority of HCI research focuses on specific use cases of LLMs within design or other domains (e.g., LLMs for creating personas).We argue for a systematic exploration of LLM capabilities to broadly identify many situations in which they could be useful in early-phase design work. To illustrate this matchmaking approach, we first curate a subset of LLM capabilities, (e.g., text summarization, paraphrasing, dialog generation), and enumerate several design activities they can support (e.g., moderating a brainstorm session, design deliberation). Next, we share a design experiment where we used chatGPT to explore if and how LLMs might moderate a brainstorming session between two designers. By reflecting on our LLM-in-the-loop ideation experience, we hope to highlight opportunities for interaction design researchers to innovate LLM-based interactions for designers.

[PDF]

New generative machine learning models provide opportunities to support design work in various parts of the design process. This study investigates how generative machine learning and large language models such as GPT-3 may play a part in co-creative design processes of ideation and early prototyping and sketching. A workshop was conducted in which design practitioners and design researchers developed design concepts for a provided design case, with the help of GPT-3. The findings point to three main themes, including the practical usefulness and limitations of the system in this context, how the form of interaction shapes users expectations on the systems capabilities, and finally, how the discourse of AI limits and enables certain aspects of co-creation. The discussion outlines some design implications and alternative framings of the co-creative design practices based on post-human perspectives on design and technology use.

[PDF]

Although studies of creativity have shown that people can generate better ideas when they collaborate with one another, it is also known that idea generation within groups often leads to incremental improvements of existing ideas rather than completely new ones. In a recent study, we investigated the use of the brainsketching idea generation technique paired with different forms of stimuli designed to prompt creativity. Using Linkography, we found that introducing such creativity stimuli into group sketching activities positively affected the creative ideation process and resulted in the generation of more novel ideas, both by individuals and by the group. 

In our study, it was a human facilitator who selected and introduced specific creativity stimuli into the idea generation process, based on their experience as design process facilitator, thereby affecting the direction that the ideation took. Building on our research, it is our proposition that this role of facilitating shared creativity through the introduction of creativity stimuli is something that could potentially be done by AI rather than by humans. AI could randomise stimuli, rapidly generate new stimuli and act as a novel collaborator in the design team. An AI facilitator could add value by by introducing different types of stimuli in different ways, then monitor their influence on the idea generation process and over time, adjust the stimuli to improve the quality and quantity of the design output.

[PDF]

In the shift to a new hybrid future of work, people’s creativity and collaborative interactions are increasingly mediated through digital tools. However, research during the pandemic showed that video conferencing stifles collaborative creativity. In contrast, artificial intelligence (AI) has recently enabled new forms of creativity. To this end, we propose a system design for an AI-enhanced meeting experience that utilizes the rich media being shared during online collaboration. The system integrates communication and task space elements into the same shared video conferencing medium— meeting elements such as remote participants’ video feeds, transcribed speech, chat, and content (e.g., notes and pictures) are reified as data. In our position paper, we expand on this concept and explore how such meeting data could serve as input to AI and how AI may be able to generate new inspirational content and spatial layouts to enhance collaborative creativity in virtual meetings.

[PDF]

Color theme exploration and construction is one of the essential ideation steps that designers perform. Conventionally, this requires designer’s training in understanding the relationships of colors, experience in manipulating groups of color, and ability to express themselves through the carefully curated collections of colors. The process of exploring and constructing involves large effort of tweaking and reuse in the color space that reflects the personal preference of the designer. Though digital color palettes have been prevailing in design software, the current tools don’t provide ideation support to address ever increasingly complicated design tasks, as well as integrated AI that facilitates exploration. We propose the design of an intelligent color explorer tool following the Generative Theory of Interaction, and describe it from the perspective of Human-Computer Partnership, and discuss how this principle could lead to better collaborative AI ideation.

[PDF]

AI is often thought of as being less creative than humans and there are biases against AI-generated creative works. These perceptions may present barriers to integrating AI into creative ideation teams. However, there is potential for AIs to be utilised in ways that leverage the stereotypes held about them to increase acceptance, as well as methods from social psychology which may be used to reduce bias against AIs. This paper therefore sets out to suggest how this may be achieved to improve acceptance of AIs in collaborative, creative teams with multiple human actors.

[PDF]

This paper evaluates collaboration mechanisms between professional designers from two workshops on inspiration, algorithms and creativity. The workshops were conducted on site during a publicly staged gallery exhibit, Design, Interrupted. To clearly communicate our visual research problem to participants we deployed a co-designed, public-facing interactive gallery exhibit, critically contextualizing algorithmic search within its analog bloodlines. This approach combined teaching and making to interrogate how algorithms are influencing image search practices in ideation. Workshop participants collaborated on both analog and algorithmic image mood boards and reflected upon the affordances and potentialities of both search methods. This paper concerns finding on how the interplay of the two workshops’ team compositions influenced their collaborations. Findings revealed organizational hierarchies and different design disciplines changed the nature of collaboration mechanisms.

[PDF]

Recent progress in artificial intelligence (AI) has led to an expansion of Human-Computer Interaction (HCI) possibilities, including the use of AI as a collaborator in creative endeavors, such as art. However, the implications of this cooperation remain unclear. To gain insights into the experience of artists, we conducted a study that investigates the dynamics of creative collaboration between professional playwrights and generative AIs. An intensive three-day workshop was conducted, where seven professional Italian playwrights integrated three creative tools based on GPT-3 (Playground, ChatGPT & Sudowrite) into their creative process to develop a play. The experience was documented through a pre-workshop focus group. The Integrated Behavioral Model was employed as a conceptual framework to examine how attitudes, personal agency, and perceived norms influence the intention of playwrights to engage in cooperative behavior with AI. This study examined the barriers and drivers of AI adoption in creative writing collaborations and found that exploring AI as a tool for ontological inquiry, understanding the perception of threat related to expertise, and ethical and political boundaries are crucial factors for the integration of AI into artistic practices.

[PDF]

Co-creation sessions have traditionally relied on low-fidelity, analogue tools and paper-based approaches, which can limit the speed, depth, and complexity of the design process. However, by leveraging emerging technologies such as Artificial Intelligence (AI), designers can improve collaboration with clients and facilitate more efficient design processes. We are interested in how AI tools can be operationalized in the design process and how they can facilitate collaboration between designers and clients. To address these challenges, we propose a human-centred AI collaboration tool called Visul, which bridges the gap between emerging technologies and co-ideation and co-creation sessions. Visul acts as an intelligent intermediary, increasing creativity and understanding without overshadowing the human participants. Our solution addresses three design challenges: (i) Facilitating the design process through AI automation, while allowing users to manifest control and creativity, (ii) Providing inspiration for exploration, and enabling iterative design processes, and (iii) Assisting in requirement understanding and definition while adapting interaction behaviour to improve collaboration. By using speech-to-text and text-to-image AI models, Visul, visually creates, a new way of design ideation and more effective human collaboration. With our tool, designers can achieve more efficient and productive collaboration, leading to enhanced value creation for clients.

[PDF]

Musical professionals who produce material for non-musical stakeholders often face communication challenges in the early ideation stage. Expressing musical ideas can be difficult, especially when domain-specific vocabulary is lacking. This position paper proposes the use of artificial intelligence to facilitate communication between stakeholders and accelerate the consensus-building process. Rather than fully or partially automating the creative process, the aim is to give more time for creativity by reducing time spent on defining the expected outcome. To demonstrate this point, the paper discusses two application scenarios for interactive music systems that are based on the authors’ research into gesture-to-sound mapping.

[PDF]

While AI-enabled technologies are changing the way we work individually, most AI-enabled technologies are not designed to support collocated collaboration. We present CollEagle; an interactive interface that facilitates collaborative note-taking. CollEagle enables users to create content and document collaborative processes in a tangible manner. We discuss how it mediates collaboration and describe interactions supported by the system and open challenges for future work.

[PDF]

I am particular interested in the planned ideation workshop since I work in two fields that I assume are relevant in this context. On the one hand, we did research to provide design principles for creativity support in co-located meetings (Herrmann, 2010). This research was focused on the role of facilitators and the modes and possibilities for employing large interactive walls. One essential question was about how electronic media can help to increase the extent of ideas generated by groups (Carell & Herrmann, 2009). We have explored methods to support creativity for the design of socio-technical processes (Herrmann et al., 2012). On the other hand, we are interested in research on Human Centered AI, in particular in the concept of hybrid intelligence systems (Dellermann et al., 2019; Herrmann, 2020) where the strength of both sides –human and AI_ are employed and promoted, not only to solve complex tasks more successfully but also to advance learning on both sides. This approach requires specific interaction modes (Herrmann, 2022) and the embedding of human-AI interaction into adequate organizational practices from a socio-technical point of view (Herrmann & Pfeiffer, 2022, 2023).

[PDF]

This paper addresses the role of AI in co-design, especially in the movement-based interaction design process. There is a possibility that AI could help foster the human-human co-design process in movement-based interaction and play other different roles in collaborative ideation and the potential roles of AI that could be used in co-design practices.

[PDF]

To design Positive Artificial Intelligence (AI), it is crucial to develop accurate models of wellbeing for AI optimization, with community-led approaches playing a vital role in this process. However, this process is laborious and may be overwhelming due to the vastness of wellbeing theory. Recent advancements in Generative AI, such as ChatGPT, suggest its potential in alleviating these difficulties and fostering collaboration in the context of Positive AI. This paper investigates the role of AI in supporting human-centered design activities for modeling, measurement, and action space development, by posing the question, “Can AI design Positive AI?”. Through exploring AI’s potential as a collaborator, we aim to better understand its capacity to contribute to the pursuit of wellbeing-focused AI systems.

[PDF]