Secondment: Naja Holten Møller
|
Institution of origin |
University of Copenhagen (UCPH) |
|---|---|
|
Host institution |
Institute of Applied Biosciences (INAB), Centre for Research & Technology, Hellas (CERTH) |
|
Initial objective |
Explore how UCPH’s Human-Centred AI expertise can strengthen the AHEAD consortium’s emerging standards for responsible AI in healthcare, bridging technical evaluation criteria with the human, organizational, and contextual factors that determine whether AI systems work responsibly in practice. |
During secondments at CERTH (Greece) in April and June 2026, Associate Professor Naja Holten Møller from UCPH (Denmark) collaborated with researchers across the AHEAD consortium to explore how Human-Centred AI (HCAI) can strengthen emerging standards for responsible AI in healthcare. The work addressed a central challenge in AI standardisation: while existing standards establish common processes and measurable criteria for evaluating properties of datasets such as accuracy, robustness, transparency, and accountability, they provide fewer mechanisms for systematically capturing the human, organisational, and contextual factors that ultimately determine whether AI systems are responsible and effective in practice.
Key achievements
Contributed to AHEAD’s work on responsible AI standards by exploring “openings” for how Human-Centred AI perspectives can complement existing technical evaluation frameworks and emerging international AI standards.
Initiated the work of identifying relevant papers for building the knowledge base for the AHEAD Observatory covering Human–AI alignment and feedback loops.
Designed and facilitated a workshop examining the values and assumptions embedded in the AHEAD LLM prototype, generating insights into how human values and organisational contexts shape AI system design and evaluation.
Initiated discussions on integrating Human-Centred AI into the AHEAD responsible AI framework to better account for context and use in practice.
Why this matters for AHEAD
While standards provide common methods for evaluating AI systems, their responsible use depends on how they are integrated into practice. By connecting Human-Centred AI with ongoing standardisation efforts, this work helps AHEAD develop governance approaches that combine technical evaluation with human values and organisational context.
Future work
Future work will focus on developing evaluation metrics that capture human and organisational dimensions of AI systems, expanding the Observatory’s evidence base, and integrating these perspectives into the iterative development and evaluation of the AHEAD LLM prototype.
Associate Professor Naja Holten Møller during the workshop at CERTH
