The COVID-19 crisis has created a sense of urgency in Europe to improve its use of AI for preventing and treating infectious diseases, but lack of access to data remains a key obstacle, according to panelists at a webinar hosted on March 31 by the Center for Data Innovation.
Various AI techniques can help fight infectious diseases more effectively, argued Emanuela Girardi, founder of Pop AI and member of the Italian Government’s high-level expert group on AI. For instance, machine learning models using intensive care data can help doctors triage patients and recommend different therapies. AI can analyze data from hospitals to optimize scheduling and the allocation of scare resources and equipment such as respirators and masks. Machine learning can also analyze epidemiology data to forecast the spread of the virus. Tracking systems can help identify suspected carriers and their contacts. Ségolène Martin, co-founder and chief executive officer at Kantify and ambassador of Women in AI, added that AI can help identify potential drugs to treat the disease and scan lungs to diagnose COVID-19. Kantify recently developed a model to predict the spread of COVID-19. The model is different from traditional epidemiology models in that it is not only using the number of infections and total country population, but also the population’s age and gender, as well as other variables such as the proximity between countries, the type of government, and the nature of quarantine measures.
Both Girardi and Martin agreed that an obstacle to deploying and using AI is the lack of access to high-quality, accurate data—one cannot fight what one cannot measure. In addition, both panelists said data sharing is a pressing issue in the context of the COVID-19 crisis. Girardi referred to an EU-level data sharing agreement as an important step in providing the right infrastructure to share data. According to Martin, the EU has the experience and creativity to spearhead such a solution, but currently, European countries are collecting data in silos and using different standards. Data quality also varies from one country to another. Better data could allow member states to better understand what works and what does not to combat the spread of the virus and make more informed decisions individually and collectively.
European researchers want to help. The Confederation of Laboratories for Artificial Intelligence Research in Europe (CLAIRE), a pan-European network which brings together scientists, technologists, and research institutes from 34 countries, recently sent a letter to offer its members’ expertise in AI to the European Commission and member states. Another option to facilitate collaboration is to create a dedicated EU-wide research center to coordinate AI activities faster.
Chivot argued that the EU can only catch up in the data economy if it fast-forwards the development of critical systems. There are various ways in which the EU could do so. First, the EU should invest in research and education. Second, relaxing regulations could help health companies get products to market faster. Third, the EU should prioritize data collection now, and not wait until it has realized its ambitions to create a common public data space, as that may take too long.
The COVID-19 crisis presents an opportunity for a live experiment on consolidating dispersed European efforts around AI. The goal should be to ensure healthcare practitioners and policymakers are able to use and analyze data necessary to respond to the pandemic. This is also an opportunity to accelerate the digital transformation of public and private organizations, spread a culture of data among Europeans, and improve data-driven policymaking.