The Role of Role: How AI teammates’ Function and Operational Context Shape Human Perceptions

This is the authors' abstract. We don't add key points for this paper.

AI systems in human-AI teams must contribute to team cognition and elicit the correct perceptions in humans to foster adoption in critical domains. While prior work describes AI teammate roles that could benefit team cognition, human perceptions of these roles are underexplored. To examine this gap, we conducted a factorial survey where participants evaluated six AI teammate roles in four operational contexts for perceived usefulness, adaptability, performance and trust. We found that AI teammate roles significantly affect human perceptions, with the Process Analyst receiving more positive evaluations and the Information Translator receiving generally negative evaluations. Although context alone did not alter perceptions, it interacted with the AI teammate role to alter perceptions of trust and usefulness. These findings extend HCI and HAT research on perceptions of AI teammates, broaden knowledge on role framing to foster positive perceptions and highlight context as a critical moderator for perceived usefulness and trust.