Human-Centred Oversight for Trustworthy AI in Health and Cancer Care

Human oversight will be essential for trustworthy AI-enabled healthcare, but assigning a professional “in the loop” is not enough. Unclear responsibilities, automation bias, limited explainability and added workload can undermine safety, trust, accountability and adoption.

 The challenge is to organise oversight so that it is effective, efficient and workable within real clinical and organisational settings. This requires dedicated Social Sciences and Humanities (SSH) research alongside technical and clinical development.

 Building on its track record in human-centred design and digital innovation, SMIT offers research on:

• How clinicians and patients understand, trust and challenge AI-supported decisions.

• How roles, responsibilities, liability and escalation procedures should be organised.

• Which explanations, interfaces, skills and training enable meaningful intervention.

• How oversight models affect workflows, workload, equity, acceptance and implementation.

Outcomes include evidence-based insights, user and organisational requirements, validated oversight models, and practical recommendations for design, governance and implementation.

This SSH contribution will be particularly relevant to proposals involving AI-enabled analysis, personalised recommendations or clinical decision support under the following calls:

• HORIZON-MISS-2027-02-CANCER-01 — Leveraging functional genomics to reveal novel targets for cancer treatment

• HORIZON-MISS-2027-02-CANCER-04 — Improving equitable health outcomes and added value for and with cancer patients

• HORIZON-MISS-2027-02-CANCER-05 — Pre-commercial procurement of affordable cancer solutions

• HORIZON-HLTH-2027-01-CARE-02 — Personalised approaches to reduce risks from adverse drug reactions

• HORIZON-HLTH-2027-02-TOOL-01-two-stage — AI-based predictive biomarkers for chronic non-communicable diseases

• HORIZON-HLTH-2027-03-TOOL-04 — Virtual Human Twins for integrated clinical decision support

• HORIZON-HLTH-2027-03-TOOL-08 — Towards Artificial General Intelligence for healthcare