Animesh Bondre, Gyuchan Thomas Jun, and Emilene Zitkus
Recent advancements in Artificial intelligence (AI) have demonstrated significant potential for identifying disease trajectories and clusters, among people living with Multiple long-term conditions (MLTC). Among these developments, AI-based Natural Language processing (NLP) applications are gaining increasing attention as they can replicate aspects of human language. This ability could enable applications to understand, extract, structure, and communicate health information and data through different interventions (Dambha-Miller, 2023). Despite their transformative potential, use of NLP applications in healthcare raises important ethical dilemmas like regarding patients’ autonomy, trust, reliability and accountability. These challenges need to be addressed proactively before NLP’s widespread development and deployment to avoid any unintended negative consequences. To address this, there is increasing interest in anticipating the socio-technical systems by engaging with future studies-based scenario generation.
However, the rapid pace of AI innovations is increasing the complexity of the healthcare systems. Alongside the technology developments there are also multiple social, political, environmental and economic transformations. This makes it difficult to be certain about how NLP applications in the healthcare systems may evolve in the future. As a result, the scenario generation methods for anticipating the socio-technical systems must account for pluralistic visions of the future that are dynamic, intertwined, and may be potentially conflicting while still capable of coexisting simultaneously.
To address this need, this research examines application of worldbuilding, a method that has gained increasing attention within future studies for anticipating pluralistic scenarios. Worldbuilding adopts a systems-oriented perspective towards scenario generation and enables researchers to create pluralistic scenarios by envisioning broader contexts like culture, future generations’ everyday life and values surrounding the technology.
Through worldbuilding workshops this research will explore alternative configurations of the socio-technical system involving applications of NLP for people with MLTC. The resulting scenarios will then be compared and analysed for surfacing ethical issues and examining factors affecting these issues. This process will enable discovering ethical value conflicts and tensions that may exist in in the system.
The intended outcome of this research is to leverage the findings from this pluralistic scenario anticipation process to contribute towards development of ethical guidelines or policies of emerging NLP applications in healthcare of people with Multiple long-term conditions (MLTC). These guidelines or policies will further support NLP developers and other healthcare stakeholders to proactively explore ethical dilemmas for responsible AI innovations in healthcare.
Keywords: systemic thinking, future studies, artificial intelligence, healthcare, people with learning disabilities, ethics