Research Agenda for Human-AI Teaming in Safety-Critical Settings
Year: 2026
Type: conference-paper
Abstract: Human-AI teaming (HAIT) is increasingly deployed in safety-critical settings, yet remains fragmented and insufficiently grounded in real-world practice. This paper synthesizes key factors shaping HAIT, including user characteristics and reliance risks, interface and output design, and the alignment between explainability and user interpretability. We highlight HAIT as a central sociotechnical challenge that requires continuous calibration and monitoring rather than one-time design fixes. We argue that effective HAIT requires a supportive Responsible AI infrastructure spanning design, deployment, and use, and outline a research agenda calling for integrative, real-world, and longitudinal studies.
Source: HAL
Authors: Bach, Tita Alissa, Mcgeorge, Dag
Institution: DNV (Norway)
Language: English
Subfield: Human Factors and Ergonomics
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7220800700