GenARate Together: Investigating Collaborative Image Generation Through Tangible Interaction

Generative artificial intelligence (GenAI) allows image creation through graph-based tools, yet these systems become complex quickly. We implemented a tangible user interface (TUI) and a tablet system to simplify image generation. We investigated how social context influences the user experience and sense of agency regarding the interaction modality. In a within-subject lab study with 24 participants, the tangible interface showed a significantly increased hedonic quality, while the tablet scored higher on pragmatic quality. Participants described the tangible interface as promoting collaboration, parallel work, and playful engagement, while valuing the tablet for efficiency. They also emphasized that supporting user agency required balancing the clarity of tangible interaction with the flexibility of text-based input methods. Finally, participants highlighted the potential of such systems to promote AI literacy for children and novices, while raising ethical concerns about misuse and the lack of recognition of original content in GenAI.