The European Commission is showing fresh signs of its commitment to data-driven innovation. In her recent speech at the new Commission’s final hearings, President Ursula von der Leyen added a new task for Thierry Breton’s portfolio as Commissioner for Internal Market: Leading a new “data strategy” to enable organizations to share and pool data securely. Vast amounts of data remain underutilized and isolated, so developing a strategy to capture the full benefits of data will help the EU’s digital economy grow and become more productive.
To achieve this goal, Commissioner Breton should focus on four key goals.
First, the EU data strategy should aim to improve and leverage existing public data, and increase the quantity and quality of datasets available to the public. To this end, the EU should accelerate digitization efforts and work with member states’ government agencies to develop shared pools of high-quality, application-specific training and validation data in key areas of public interest, such as agriculture, education, health care, public safety, and transportation. For example, as public health authorities provide most of the health care in Europe, the EU has an opportunity to amass extensive datasets on patients and outcomes that can fuel the development of health-related AI. The EU should foster the further development of open data policies and tools, such as the European Data Portal, which has a wide variance of participation from member states. More comprehensive data collection and higher quality public sector datasets will lead to more opportunities for companies, non-profits, and individuals to create innovative applications.
Second, the EU data strategy should streamline the adoption of common standards by all member states to accelerate data access, facilitate interoperability, and maximize the use and reuse of data between government authorities, researchers, and companies. Unfortunately, various EU information systems lack the common standards that allow for even basic interoperability. For instance, the ability to access and share medical data across the EU varies greatly, limiting the ability to train AI systems on cross-border data. In addition, high resolution geospatial data is very useful to the development of technologies, such as AI systems enabling autonomous vehicles or precision farming for land management—but only a handful of EU countries currently provide open and free access to datasets. Those that do exist are often either outdated or cannot be assembled because they are not available in similar quality and resolution. Fragmented and incomplete repositories make information difficult to analyze and compare, and member states would benefit from common data standards.
Third, the EU data strategy should encourage businesses, public sector, and non-profit organizations to open access to more of the data they hold by facilitating data sharing between companies when it is mutually beneficial. For example, many major pharmaceutical companies have begun sharing historical clinical trial data with outside researchers, including competitors, which can use this data to accelerate drug development, better understand diseases, and design more efficient clinical trials. The EU data strategy should consult industry about the conditions under which proprietary and sensitive data should be shared with third parties, as well as the most effective mechanisms to safeguard this data. For example, the EU may work with industry to develop sector-specific data trusts to encourage businesses to collect and share data responsibly.
Fourth, the strategy should improve the EU data governance framework. The EU should ensure all its member states appoint a chief data officer (CDO). It should then establish a new, independent, EU-wide advisory panel, made up of each member state’s CDO, charged with counseling the EU on how to maximize opportunities to innovate with data, including with AI and the Internet of Things. Finally, EU policymakers should amend the General Data Protection Regulation (GDPR), which they developed before fully understanding AI and which has put unnecessary constraints on how European businesses collect and use data. Some of its requirements, such as to minimize data collection or retain data for limited periods of time, negatively affect the amount of data available to organizations to train and use AI systems.
Overall, the right data strategy for the EU should improve and increase data made publicly available, and support its broader circulation to all organizations in the digital economy. It should also encourage responsible and secure data sharing across organizations, facilitate data collection, and incentivize the use of data to create new business opportunities and societal benefits.
Image credits: Wikimedia Commons.