Recap Design for Values in AI for Medicine and Healthcare

The workshop entitled “Design for Values in AI for Medicine and Healthcare” has been hosted in Delft on November 25th, 2022 as the result of a collaboration between the Delft Digital Ethics Center and the World Health Organization.

Uses of AI-based systems in morally salient contexts such as medicine and healthcare can bring about epistemic, societal, and ethical concerns that need to be timely tackled. Which values do AI systems need to uphold for their ethical implementation in medicine? How can we ensure their societal acceptance? And how can these values be effectively operationalized? The overarching aim of the workshop was to bring together experts from relevant disciplines (computer science, ethics, healthcare organizations, and medical practitioners) to discuss challenges and promises of a Design for Values approach in medical AI from different angles. Bringing together different stakeholders in a highly interdisciplinary setting facilitated fruitful exchanges and lively discussions. 

The event started with a presentation by Jeroen van den Hoven (TU Delft) on the Design for Values approach and its relevance in crucial fields such as medicine and healthcare. He stressed how conceptual clarity regarding the requirements these technologies should meet is of paramount importance to their ethically sound development and to foster social acceptability.
Simão Campos, from the ITU-WHO AI for Health Focus Group, described their overall objectives, emphasizing the potential of AI systems to enable healthcare delivery making it available worldwide. Among the main functions of the focus group, he highlighted enabling governance and data sharing, facilitating cooperation, and implementing sustainable models.
Afterwards, Francesca Pratesi from the research infrastructure (RI) SoBigData++ shared her insights from the perspective of computer science. She presented the main goals of the RI, particularly with reference to case studies in the medical setting focusing on the methods and datasets developed within the SoBigData++ project.

Eva Weicken, from the Fraunhofer Heinrich Hertz Institute for Telecommunications in Berlin and member of the ITU/WHO focus group talked about standardization in health AI and the need of ongoing exchange in an interdisciplinary environment toward healthcare provision and regulation supported by AI. The presentation by Alessandro Bozzon (TU Delft) focused on the operationalization of conflicting values in AI-enabled healthcare, emphasizing how design research and methodologies and human-computer interactions are key to address these challenges. Finally, Marc Lecoultre, also from the ITU-WHO AI for Health Focus Group, focused on the AI4H Assessment Platform, a software platform to support the assessment of AI algorithms globally.
The event was concluded with a discussion and questions from the audience attending the workshop in person and online.