Artificial intelligence systems that record voice and video during pediatric emergencies are
emerging as human-computer interaction (HCI) technologies with direct implications for clinical work,
promising improvements in documentation, team performance, and post-event debriefing. Yet the
perspectives of those most affected, including clinicians, parents, and child patients, remain largely
absent from the design and governance of these technologies. This position paper argues that this has
direct consequences for the legitimacy and effectiveness of these systems. We examine four areas where
these missing perspectives prove consequential (consent, emotional impact, surveillance dynamics, and
participatory governance) and propose four positions for reorienting AI recording in pediatric emergency
care toward stakeholder-centered HCI inquiry.