France leads push for widespread adoption of European AI

“We have to stop being naïve.” That was French President Emmanuel Macron’s message as he called on public authorities and industry to support European AI systems, during the Adopt AI Summit in Paris on November 25.

The event, billed by organisers as the Davos for AI, in reference to the World Economic Forum, is the latest signal that France wants to lead efforts to develop homegrown AI models instead of relying on US and Chinese solutions.

It follows the AI Action Summit in Paris in February, where €109 billion in investments for infrastructure projects in France were announced, and the digital sovereignty summit co-organised by France and Germany in Berlin on November 18.

In Berlin, German Chancellor Friedrich Merz called for “radical simplification” of the European regulatory landscape. This was also one of three priorities Macron outlined for action at the European level during this week’s event at Paris’s Grand Palais.

“Simplification is key for all of us if we want, in a certain way, to dismantle the tariffs we inflicted on our own economies,” he said. “The best simplification is the creation of the 28th regime for innovative companies and the implementation in real terms of our single digital market.”

The European Commission will present a proposal early next year for the 28th regime, an optional EU-wide corporate legal framework making it easier for companies and investors to operate in different EU countries.

Accelerating investments in innovation was the second principle put forward by Macron. France and Germany have agreed to work towards a new Important Project of Common European Interest (IPCEI) on AI, which he said would “promote the financing of high-impact research, development and first industrial deployment projects in AI technologies.” IPCEIs are an EU tool allowing member states to channel resources to industry-led projects without running foul of state aid rules.

Innovation in AI must also mean investing “much more” in quantum and chip technologies, Macron said. He pledged to “push for more European investment at the aggregated level in our budgets,” and to support the development of the Capital Markets Union to mobilise more private investment.

The final principle is protection. Macron called for sovereignty criteria to be included in the EU’s selection of AI gigafactories and for an EU preference in public procurement. “In China, you have a Chinese exclusivity; in the US, you have a US preference,” he said. “The European Union is the only place in the world where you have a non-European preference de facto.”

He also urged large companies to play their part. “When there is an existing, competitive European solution, please adopt it,” he said.

The French and German governments want to lead by example. At their joint summit last week, they announced their intention to establish a new public-private partnership with French AI start-up Mistral and German software company SAP to deploy AI solutions to enhance public services.

Growth potential

Earlier in the day, French economist Philippe Aghion, who co-chaired the French AI commission that last year presented a series of recommendations to the government, said he was “cautiously optimistic” about the potential of AI to drive economic growth.

“We believe that there is enormous potential, but we need the right policies and institutions to steer the technology in the right direction,” he told the summit. “I am very optimistic on the technology. I am a bit less optimistic on the institutions.”

The growth potential lies in AI’s ability to automate not only the production of goods and services, but also tasks in the production of ideas, he said. While AI will replace certain jobs, it could also make firms more competitive and create jobs by making it easier to find new ideas.

“Very often, new ideas are combinations of old ideas, but with AI you can combine much more, so you have many more possibilities, and then AI helps you select from all those possibilities much faster.”

Related articles

• The EU’s AI in science plan gets a mixed reception
• State of AI 2025: five key charts for Europeans
• EU launches AI strategies to boost competitiveness and science

But growth is by no means guaranteed. One of the big risks, according to Aghion, co-recipient of this year’s Nobel Prize for economics for his work on the links between innovation and growth, is that US giants will dominate AI, as they did with cloud services.

Effective competition policy will be crucial, including supporting open-source solutions and considering the impact of mergers and acquisitions on future innovation. On the other hand, it’s important to avoid over-regulating, as large incumbents know how to navigate the regulations, unlike new entrants, he said.

He also praised the Danish “flexicurity” system that makes it relatively easy for firms to restructure in case of failure but offers generous unemployment benefits and retraining opportunities.

Adoption targets

As the name of the summit suggests, its aim was to bring together industry, start-ups and the public sector to accelerate AI adoption. In July, France adopted the Osez l'IA (Dare to do AI) plan, which wants to see 100% of large companies, 80% of SMEs and 50% of very small businesses using AI by 2030.

“We need to be better in the downstream segment of AI: AI-enhanced cloud services, solutions, applications, et cetera,” Macron said.

“The US and China are at the forefront today of innovation, [large language model] capacities and so on. But if we accelerate in terms of adoption, we will help European champions, and we will attract new champions.” AI adoption is also “the best way to fertilise the rest of the economy in terms of productivity,” he said.

Anne Bouverot, Macron’s special envoy for AI and co-chair of the AI commission with Aghion, said successful adoption depends on trust. “Our fears about AI used to be broad and almost science fiction-like [. . .] Now, they’re more focused.” Concerns are primarily related to democracy and disinformation, and jobs, she said.

In Europe, individual adoption of AI is high in people’s personal lives and for private uses at work, known as shadow AI, but general adoption by companies is more difficult, she said. “Scaling requires training, it requires social dialogue, it requires strong leadership engagement of company chief executives, and it also requires taking care of people whose jobs will change.”