[LUM#22] Governing data
Artificial intelligence relies on the availability of data in sufficient quantity and quality. But how can we ensure that research data is made available in compliance with the law? This is the aim of the mission entrusted to Agnès Robin by the French Ministry of Higher Education and Research Research.

"In a scientific context, whether it's AI research or anything else, the challenge is to clearly identify the legislative and regulatory framework applicable to the data you want to reuse," explains Agnès Robin, teacher-researcher at the Innovation, Communication and Market Laboratory (Licem*). While the principle of opening up research data has been affirmed since 2012, it does have its limits. The first is justified by the need to protect personal data. A historian digging through archives, for example, may wish to exploit civil status data. " The researcher will only be able to disseminate the data resulting from his or her research once it has been anonymized," explains the legal expert.
Legal resources platform
In the case of data protected by copyright or sui generis database law, he will need to obtain the agreement of the rights holders, "except in the case of text and data mining, since this operation was the subject of an exception in 2019 ". Finally, data covered by secrecy (defense secrets, medical secrets, business secrets, etc.) "are absolutely protected and cannot be made available, except through a trusted third party".
These three types of constraint have different effects, and require very close attention. So, "the aim of my mission is to provide researchers and research support services with a platform to help them analyze the data sets they use or produce, by providing them with precise legal information", explains Agnès Robin.
Data sharing
While the open science policy clearly aims to make as much data as possible available to research, it also requires that this be done in strict compliance with the rules : "as open as possible, no more closed than necessary. At the root of open science is the idea that data shouldn't just sit on hard drives, when it could be used by others to advance research", explains the researcher.
To ensure this sharing, the European Union has set up a common data space for open science, the European open science cloud. Health data, for their part, are subject to specific regulations still in the process of being adopted, which will eventually enable them to be shared via theEuropean health data space. " The stakes are enormous, both for research and for patients," concludes Agnès Robin.
AI law
On May 17, 2024, the European Union adopted a regulation designed to guarantee AI that respects fundamental rights. "This is the very first regulation on AI in the world," stresses Agnès Robin. Before, there was nothing!" In particular, it introduces the liability that can result from the implementation of generative AI systems. "In other words, we need to be able to explain how AI analyzes data, what data the analysis is based on, and what rules may be used to make decisions that concern individuals. Which is far from always being the case.
UM podcasts are now available on your favorite platforms (Spotify, Deezer, Apple podcasts, Amazon Music...).