Two years ago, ChatGPT was virtually unknown. Generative AI became an instant hit. Today, two out of three French people use chatbot-type conversational agents. Just as businesses rushed to embrace the Internet 30 years ago, companies are jumping headlong into the AI race to stay ahead. While six in ten North American companies have adopted AI, fewer than one in ten in Europe have followed suit.
By François Lhemery, French Foreign Trade Advisor (CCE) and Lawyer at SAMMAN Law & Corporate Affairs
The AI Act: A Pioneering European Regulation
The European regulation on AI (AI Act) was adopted on May 21st, 2024. One of its objectives is to regulate the use of this technology in order to promote the development of trustworthy AI, through a risk-based approach. The regulation sets out tiered obligations for stakeholders, depending on the level of risk (minimal or none/limited/high/unacceptable) that specific use cases may pose.
Much like the General Data Protection Regulation (GDPR) before it, the AI Act sets a standard for AI regulation. It will certainly influence countries that are preparing to legislate in this area, such as the United Kingdom, Canada and Japan, all of which have previously aligned with the GDPR.
A Foundation Act
The Executive Vice-President for Tech Sovereignty, Security and Democracy, Henna Virkkunen, has outlined her key priorities:
- Developing implementing acts and guidelines during the phased rollout of the AI Act by August 2nd, 2026
- Defining a strategy for AI adoption to accelerate the deployment of AI in industry and public administration to ensure Europe’s future competitiveness
Launching AI Factories to ensure that startups and the AI industry have access to high-performance computing (HPC) to drive innovation
Three key challenges Europe must tackle now
The first challenge is energy. At the end of the summer Microsoft announced that it would restart the No. 1 reactor at the Three Mile Island nuclear power plant to power its U.S. data centers. According to Microsoft President Brad Smith, the company’s power consumption will increase sixfold between 2020 and 2030. This figure is in line with the 2024 report of the International Energy Agency, which forecasts that annual data center electricity consumption will double between 2022 and 2026 equaling Japan’s electricity consumption. Meanwhile, Google has just signed an agreement with the startup Kairos Power to build six small modular reactors (SMRs) to meet its energy needs. Can Europe ignore this new reality and pursue the Green Deal goal set five years ago of reducing greenhouse gas emissions by 55% by 2030?
AI is also putting democracy to the test. In its 2024 Global Risk Report, the World Economic Forum identified disinformation as the top near-term global risk. The website NewsGuard has identified more than 1,100 unreliable AI-generated news sites. Remember the bedbug psychosis in France in the fall of 2023, a campaign orchestrated by Russia. The impact of the manipulation of information and its dissemination on social media using AI-powered tools, came to light on December 6th, when the Constitutional Court of Romania decided to cancel the presidential election process. Major AI players are implementing filters and measures against disinformation, with the notable exception of Grok, the chatbot developed by xAI, a startup founded by Elon Musk.
Finally, the technological race and the need for gigantic amounts of capital are putting our sovereignty under pressure. Last quarter, OpenAI and xAI closed record fundraising rounds of $6 billion. Microsoft, for its part, announced on January 3rd that it will invest $80 billion in AI data centers by 2025. This illustrates the huge amounts of money the players are spending. Consolidation is likely to take place in the next 3 to 5 years, while no European player is currently in the top 5 worldwide.
In this context, will the use of chatbots still be free in 5 years, given that the development of general-purpose AI costs billions and a query on ChatGPT consumes 10 times more energy than a query on Google?
How will we be able to guarantee and finance the pluralism of information available on the open internet when AI can create “information bubbles”?
As the new Commission tries to curb Europe’s economic decline, shouldn’t it also acknowledge its technological dependencies and invest massively and immediately in the infrastructures of tomorrow (e.g. energy, telecom networks and HPC) to create the conditions for technological leadership by 2040?
KEY TAKEAWAYS
- The technological revolution of AI continues at a relentless pace, led by the American tech giants who are investing colossal amounts of cash;
- Europe must ensure that its regulatory framework allows its players to innovate and become tomorrow’s champions, by tackling the three major challenges of AI.
