{"id":1171,"date":"2026-05-15T07:00:07","date_gmt":"2026-05-15T07:00:07","guid":{"rendered":"https:\/\/blogs.bmj.com\/bmjleader\/?p=1171"},"modified":"2026-05-06T12:08:50","modified_gmt":"2026-05-06T12:08:50","slug":"aligning-trust-across-the-healthcare-workforce-a-missing-ingredient-in-ai-adoption-by-karen-wallace","status":"publish","type":"post","link":"https:\/\/blogs.bmj.com\/bmjleader\/2026\/05\/15\/aligning-trust-across-the-healthcare-workforce-a-missing-ingredient-in-ai-adoption-by-karen-wallace\/","title":{"rendered":"Aligning trust across the healthcare workforce: a missing ingredient in AI adoption. By Karen Wallace"},"content":{"rendered":"

Trust is widely described as essential for the safe and responsible deployment of artificial intelligence in healthcare. Yet trust is not a single, uniform construct. Clinicians, operational teams, and organisational leaders encounter AI through different forms of professional responsibility, risk, and accountability. National frameworks such as the <\/span>NHS AI Playbook<\/span><\/a> and <\/span>FUTURE-AI<\/span><\/a> emphasise transparency, fairness, and ethical deployment, but they operate largely at system level. Far less attention is given to how these principles translate into the lived realities of healthcare work.<\/span><\/p>\n

A mixed\u2011methods study involving 218 clinicians, operational colleagues, and leaders in an independent healthcare provider underscores this point. Quantitative differences between roles were statistically modest, but a striking pattern emerged: more than 60% of respondents across all groups selected \u201cMaybe\u201d when asked whether their organisation was prepared for AI. This finding mirrors UK and international studies in which uncertainty, not resistance, is the dominant early\u2011stage response to AI adoption. Uncertainty is a normal, evidence\u2011based feature of responsible digital transformation, signalling neither rejection nor organisational deficit.<\/span><\/p>\n

Trust in ethical implementation showed a similarly moderate pattern, averaging 6.6\/10 across respondents. Again, this aligns with broader research demonstrating that staff express provisional trust until governance becomes visible at the point of use.\u00b9 Early\u2011stage confidence in AI typically sits in the mid-range because professionals require clear assurance on safety, accountability, and workflow integration before committing to full adoption.<\/span><\/p>\n

While the quantitative data suggest broad similarity across roles, the qualitative analysis revealed meaningful interpretive differences that matter for leaders. Clinicians consistently described trust in terms of <\/span>defendability<\/span><\/i>: the ability to justify an AI\u2011supported decision within medico\u2011legal and professional standards of care \u2014 a concern <\/span>increasingly recognised in UK practice<\/span><\/a>. Operational teams emphasised <\/span>procedural clarity<\/span><\/i>, highlighting the importance of workflow ownership, escalation routes, and integration into existing operational processes. Leaders, meanwhile, emphasised <\/span>governance visibility<\/span><\/i> and regulatory assurance, reflecting organisational accountability for risk.<\/span><\/p>\n

These perspectives are not contradictory; they represent different expressions of the same underlying requirement: trust must be experienced in ways that resonate with each role\u2019s exposure to risk and accountability. When these expectations are not aligned, AI adoption becomes slower, not because the workforce rejects innovation, but because the conditions for confidence are unclear.<\/span><\/p>\n

The findings show that organisational readiness for AI should be understood not as a technical state but as a relational construct. Staff do not gain confidence from accuracy statistics alone; they gain confidence when governance is visible, decision pathways are transparent, and role-specific responsibilities are unambiguous. High-level principles including fairness, explainability, oversight, must be translated into actionable structures that staff can observe, use, and rely upon in real practice.<\/span><\/p>\n

Across clinical, operational, and leadership groups, one theme consistently ranked as the most important enabler of adoption: ethical governance. More than 90% of respondents agreed that colleagues across roles should have a meaningful voice in decisions about AI implementation. Yet qualitative responses revealed uncertainty about whether this inclusion actually occurs. <\/span>This gap between formal commitment and lived experience is well documented in AI governance literature<\/span><\/a> and represents one of the most significant barriers to workforce confidence.<\/span><\/p>\n

To bridge this gap, and guided directly by these findings, I developed a role\u2011sensitive organisational model for implementation: the CORE approach \u2014 Confidence, Organisational Readiness, Role Alignment, and Ethical Governance. Rather than a generic framework, it was designed explicitly with different professional groups in mind and grounded in the empirical patterns of this study.<\/span><\/p>\n