In the second of a series of blogs from our global offices, we provide a overview of key trends in artificial intelligence in France.

What is France’s strategy for Artificial Intelligence?

The French president, Emmanuel Macron, announced in March 2018 his ambition for France to become a global leader of the artificial intelligence (AI) ecosystem.

The first phase of the National Programme included an initial investment of €1.5 billion into the creation of a network of interdisciplinary institutes dedicated to artificial intelligence (the “3IA” institutes) and the financing of multiple AI projects overseen by Bpifrance.

The second phase will provide for €2 billion of private and public funding to attract and train new talent. More than half of this financial effort will be allocated to implement a network of global academic centres of excellence and a massive plan for AI within universities.

The goal is to promote future French and European AI champions and improve the competitiveness of existing companies. The measures envisioned by 2025 are:

  • Investment in new generation on-board (i.e. embedded) AI systems and platforms for software development, including open-source, (€1.22 billion);
  • Supporting 500 small and medium sized companies with the introduction of AI solutions (€25 million);
  • Supporting existing business incubators to triple the number of AI specialized start-ups by 2025 (€40 million); and
  • Launching proof-of-solution technologies for more frugal use of energy in AI in order to attain climate targets (€120 million).

This strategy is synchronized with the European plan for artificial intelligence, which aims to develop a leading offer in terms of embedded AI and the use of AI to fast forward environmental transition.

More information on the French National Strategy for AI can be found here. A detailed “mission” statement was published under the aegis of the French Parliament. It is available in English as “For a Meaningful Artificial Intelligence: Towards a French and European Strategy“.

What are the leading French research institutions for Artificial Intelligence?

To enable the French artificial intelligence ecosystem to develop and to accelerate its growth, the National Strategy relies in particular on a network of Interdisciplinary Institutes for artificial intelligence (3IA institutes), for which the following four projects have been selected: Grenoble “MIAI Grenoble Alpes”, Nice-Sophia Antipolis “3IA Côte d’Azur”, Paris “PR[AI]RIE”, and Toulouse “ANITI”.

These four institutes aim to conduct research on AI, offer training for students and professionals, support companies (large, medium or start-up) in their activities related to AI and, more generally, publish articles to inform citizens about AI.

Driven by the same ambition, other research centres of excellence have been created in prestigious French schools such as the Data IA Institute (Paris Saclay University), SCAI (Sorbonne University), and Hi! Paris (HEC Paris and Institut Polytechnique de Paris).

For example, the Department of Computer Science, Data and Artificial Intelligence at the Institut Polytechnique de Paris, aims to:

  • contribute to the foundations of computer science;
  • develop powerful algorithms to provide the basis for an efficient, reliable and secure digital world;
  • develop algorithms and statistical processing methods for learning, optimization and data analysis; and
  • explore and roll out new AI applications for industry and society.

What laws and regulations is France developing for Artificial Intelligence?

France is aligned with the European Union’s strategy for data and policy to ensure the development of “human-centred” artificial intelligence. On 19 February 2020, the European Commission adopted a White Paper on Artificial Intelligence: a European approach to excellence and trust.

This White Paper has led to a proposal for a European Regulation laying down harmonised rules on artificial intelligence and amending certain EU legislative acts in order to establish a harmonised definition of artificial intelligence and define the issues at stake, in particular by adopting a risk-based approach to liability (commonly known as the “AI Act“).

On 18 June 2021 the French Commission Nationale de l’Informatique et des Libertés (CNIL), data protection authorities in other Member States and the European Data Protection Supervisor adopted an opinion on the AI Act. They indicated, in particular, the need to broaden the scope of prohibited AI systems and to clarify their definition. The CNIL and the other Member States’ authorities asked that the governance of the “European Committee on Artificial Intelligence” (CEIA) be clarified, both to guarantee its independence and to strengthen its powers to enable it to exercise real control, especially when AI systems are implemented on a European scale.

At the national level in France, a constitutional bill on the Charter of Artificial Intelligence and Algorithms was submitted on 15 January 2020 in the National Assembly. The aim is to include a reference to the charter in the preamble of the Constitution of France. The proposal is to enshrine in constitutional law that an AI cannot have legal personality. The notion of artificial intelligence is understood in the charter as “an algorithm that evolves in its structure and learns beyond its initial programming”. It sets out principles that AI must respect (such as respecting human orders) and includes requirements for audits and monitoring the evolution of AI towards decision-making autonomy. However, the proposal has not been incorporated into the Constitution and no longer seems to be under consideration.

The law of 2 August 2021 relating to bioethics requires healthcare professionals to inform patients when they use medical devices based on machine learning (“algorithmic data processing whose learning has been carried out on the basis of massive data”) and, where appropriate, to explain the resulting medical interpretation.

AI is expected to be important in strategic sectors, including transport, agriculture, construction and information and communication technologies. To encourage investment a guarantee is needed: that intellectual property law will robustly and reliably protect inventions. A key debate is whether to protect inventions by AI (e.g. by patents) and works by AI (e.g. under copyright). Under French IP law, the inventor for the purpose of patent protection, must be a natural person (for example, the inventor cannot be a company). Now the European Patent Office (EPO) decisions in the DABUS applications (see here) have definitively ruled out patent protection for inventions by AI. As for copyright, in order for a work to be protectable in France, it must bear the imprint of the personality of its author; this appears to require a natural person and rules out protection for works generated solely by AI.

Which are the leading French companies for Artificial Intelligence?

France is home to significant AI companies including six “unicorns”:

  • Alan;
  • ContentSquare;
  • Dental Monitoring;
  • Meero;
  • Mirakl; and
  • Shift Technology.

More than 500 AI specialised start-ups were reported in 2021, which corresponds to an 11% increase compared to 2020.