France released its AI strategy. What does it mean for Europe? 🤖
Ethics could be the core principle of a European model of Artifical Intelligence
On Thursday March 28th, Cedric Villani, an MP and renown mathematician sometimes nicknamed “The Lady Gaga of Mathematics” as it has been recently recalled by The Verge, released his report on Artifical Intelligence born out of a 6-month mission requested by the Prime Minister Edouard Philippe.
The main ideas of the 235-page long document entitled “For a meaningful Artificial Intelligence. Towards a French and European strategy” were touched upon during the “AI for Humanity” Summit held last Thursday in Paris, during which the President Emmanuel Macron announced his views for the French public strategy in terms of AI. The whole sequence was closed by a presidential interview about the stakes of this promising technology in the prestigious US tech magazine WIRED, where Emmanuel Macron called for a European model of AI.
All of these materials are available in English: a shortened version of the full report (150p.) and its synthesis (10 p.), as well as Emmanuel Macron’s closing speech (1h) and of course his interwiew to WIRED.
What do all of those announcements mean for Europe?
Overall, most of the Villani report concerns the French context: how to make French research more attractive to worldwide talents, how to better connect various AI actors throughout the territory, etc.
But its does include several interesting notions for our European partners: firstly with several suggestions of measures that aim to develop an AI industry at the scale of the continent; then by underlining what a French or European way of doing AI could look like. A large part of the document focuses indeed on the societal and environmental impacts of AI.
Building a European industry of data and AI
Out of the 6 main chapters of the report, the first one, called “An Economic Policy Based on Data”, is the one that directly recommends some European-scale policies.
Its first plea is to “reinforce the European data ecosystem”. Several schemes are being considered: incentivizing economic actors into pooling their data in order to create a larger data market in Europe is among them; the possibility to access data held by private entities in some determined cases is mentioned as well. The report reiterates the right for citizens to obtain portability of their data, recognized by the GDPR and to a further extent by the Law for a Digital Republic in France (2016). Regarding the free flow of data, the Villani mission calls for a strong legal control of transfers outside of the EU borders. The last point underlined is the need for the European legal framework to evolve so as to allow and promote new kinds of data uses such as the practices of text an data mining (TDM).
The second angle is to initiate a European-scale AI industry: “Europe has everything it needs to become a leading player in the global AI race”. In order to do so, several measures are listed: developing European robotics through a French-German-Italian leading partnership, making AI for transportation one of the priorities of a future European DARPA-like agency for innovation, supporting the components industry (CPUs, GPUs) where Europe has not a full control of the supply chain, and setting up AI infrastructures in Europe (cloud providers…).
As for the French economy, the report recommends to stress the efforts in 4 “strategic sectors” where it considers the country stands a chance to compete on a global level: Healthcare, the Environment, Transport-Mobility and Defense-Security.
Read more in the report (p.20 and p.49)
Europe and an Ethical view of AI
Out of the 6 chapters of the Villani report, half are fully dedicated to the idea of an ethical AI.
Three main angles are detailed: using AI for the ecological transition, ensuring AIs are not “black boxes” out of control, and promoting inclusive and diverse AI ecosystems.
Although these parts are filled with references to the French context, the stakes go far beyond the French borders and the report drafts potential principles on which a European model of AI could be based.
Regarding the green economy for example, the synthesis emphasized that “France and Europe can spearhead this smart ecological transition by raising awareness on the international arena. The primary task is to consider both the impact of AI on achievement of the UN’s sustainable development goals, how it puts pressure on certain goals and how it can accelerate others”. As for diversity, the statement that “Artificial intelligence must not become a new way of excluding parts of the population” is undoubtedly meant to be a rule across the whole continent. The same way, calls to “open the black box”, “implement ethics by design” or “set up and AI ethics committee” would barely make any sense if not put in the context of a Europe-wide market.
Stressing this much on ethics, rather than purely on economic policies, did raise criticism among some economical players. Yet this approach was deliberately chosen in order to suggest a way through which Europe could emerge as an alternative model to the 2 current leaders of the tech economy: the US and China.
These views have largely been embraced by the President Macron in his speech and interview. Beyond the economic measures, such as a €1.5 billion funding over the next 5 years and announcements of new AI investments in Paris from major tech companies, it is probably the potential role of Europe as a global reference that is the most outstanding when reading the discussion in WIRED:
“The key driver should not only be technological progress, but human progress. This is a huge issue. I do believe that Europe is a place where we are able to assert collective preferences and articulate them with universal values. I mean, Europe is the place where the DNA of democracy was shaped, and therefore I think Europe has to get to grips with what could become a big challenge for democracies” — Emmanuel Macron in WIRED
In his closing speech, the President advocated for several propositions in that regard, such as setting up an IPCC-like international group of experts for AI, or initiating a European reflection on algorithms control. On the French level, one of its core announcements is that all algorithms created by the government or by companies subsidized by public money will have to be opened. These measures, and many more, will take shape in the forthcoming months.
