Responsible Artificial Intelligence Off-Boarding in Radiology: Staff Perspectives on Decommissioning and a Proposed Withdrawal Framework.

Abstract

Artificial intelligence (AI) governance commonly emphasizes procurement, validation, deployment, and performance monitoring but lack guidance on how embedded AI tools should be withdrawn when funding, contracts, or strategy change abruptly. We examined staff experience after withdrawal of a chest radiograph AI triage tool and developed a practical framework for responsible AI off-boarding. An anonymous staff survey was circulated 2 months after decommissioning of a chest radiograph AI triage tool at a multisite National Health Service trust. The survey repeated selected items from three earlier implementation-phase surveys, with added decommissioning-specific questions on workflow, perceived patient benefit, emotional burden, and future AI engagement. Quantitative responses were summarized descriptively. Free text responses were analyzed deductively and interpreted using the Job Demands-Resources model. The response rate was 21.4% (40 of 187), comparable to earlier survey rounds. Perceived patient benefit from AI remained stable, with 70% (28 of 40) agreement postdecommissioning versus 71.1% (32 of 45) pre-implementation, 65.5% (19 of 29) early implementation, and 67.9% (36 of 53) late implementation. Perceived logistical burden postdecommissioning (35%, 14 of 40) was higher than at late implementation (26.4%, 14 of 53) but lower than at early implementation (51.7%, 15 of 29). Response rate from reporting staff was low (5 of 40), with two voicing disappointment in AI tool withdrawal, and one stating they had grown reliant on the tool. Across 22 free text entries, frequent themes included loss of clinical value or pathway efficiency, operational relief after withdrawal, and patient-facing emotional labor during AI-enabled escalation. Decommissioning produces both operational relief and perceived clinical loss, with effects differing by staff role. We propose a three-phase AI Off-Boarding Protocol: prewithdrawal assessment, graduated transition and postwithdrawal support. Decommissioning management should be treated as a central part of responsible governance across the AI life cycle.