HealthDARES kicks off to a great start!
A three day expert-led course focused on responsible innovation, ethics and regulatory strategies
On Monday morning, 44 participants from diverse backgrounds – academia, the private sector, and policymakers – convened at ETH Zurich for the 2025 edition of HealthDares, an advanced course aiming to foster a critical understanding of the ethical and regulatory challenges arising from the increasing integration of big data and artificial intelligence (AI) within the healthcare sector.
After coffee and refreshments, the morning session began with foundational knowledge and concepts of the health data ecosystem and its stakeholders. Prof. Vayena (ETH Zurich) and Elettra Ronchi (WHO, Europe Consultant) introduced the emergence of the big data health ecosystem. They answered questions like “What is the health data ecosystem and how does the data lifecycle work? Who are the stakeholders involved and what are their roles? Which specific concerns are associated with unconventional sources of health data?”
Katrin Crameri, from the Federal Office of Public Health FOPH then discussed Health data and digitalization in Switzerland. She provided insights into current initiatives related to data sharing and digitalization and identified key barriers within the Swiss data ecosystem (e.g., technological interoperability, collaboration among stakeholders, digital literacy, etc.). As she stated:
“It is not ok that we know how many Mercedes-Benz cars drive on swiss roads, but not how many diabetes patients there are”Katrin Crameri
The afternoon then shifted focus to look at the foundations of artificial intelligence in health. Topics examined included algorithmic design, functionality and applications in health and medicine, the role of Large Language Models (LLMs), the impact of biased datasets, and privacy-enhancing technology.
Dr. Michael Moor (ETH Zurich) began this session with an introduction to AI and data science for health, describing algorithmic design and machine learning, and how AI and machine learning works. His talk raised and addressed the question of what is particular about large language models such as ChatGPT and what are current typical applications of AI?
Prof. Mennatallah El-Assady from the ETH AI Center followed this introduction with a closer look biased datasets and AI visualization. Not only did her presentation provide examples of biased datasets but she identified possible solutions of using visualizations to explain and understand black box AI models.
The last presentation of the day was from Prof. Jean-Pierre Hubaux (EPFL and Tune Insight SA), who talked about privacy-enhancing technologies for personalized health. Prof. Hubaux began by defining Privacy-Enhancing Technologies (PETs) and then explained how PETs can enhance access and sharing of health data while protecting individual privacy. Yet, he also detailed how PETs themselves also bring challenges.
The day concluded with an interactive case study analysis. Participants split into groups to analyse a case study and then came back together to compare their solutions to the ethical dilemma in a plenary discussion. This format fostered lively discussions and collaborative problem-solving among participants. These discussions then continued at the Apero and networking event held at the impressive Dozentenfoyer, ETH Main building.
