In order to compete for the European AI crown, Britain and France compete to make moves

Original source: Financial Breakfast

Image source: Generated by Unbounded AI‌

AI is seen as a revolutionary technology and therefore of great strategic importance to governments around the world. The enthusiasm for the technology has been sparked in part by the virality of the Microsoft-backed Open Artificial Intelligence (OpenAI) ChatGPT project.

So, who will lead the race for the European AI crown? At present, it seems that Britain and France are vying for the status of Europe's artificial intelligence capital.

French President Emmanuel Macron and British Prime Minister Rishi Sunak have both made bold statements about artificial intelligence in recent weeks, as both seek to gain a piece of the closely watched market.

The first is that France has frequently brushed up its presence on this issue recently. On June 18 this year, Macron said at Viva Tech, the annual French technology conference, "I think we are the number one power in the European continent (in the field of artificial intelligence), and we must accelerate our development."

**France: The hope of the whole village? **

The ambition is set, where does the money come from? At the Viva Tech conference in Paris, Macron announced 500 million euros ($562 million) in new funding to create new artificial intelligence "champions." This is in addition to the government's previous commitments ** including a pledge to spend 1.5 billion euros in artificial intelligence by 2022 to catch up with the US and Chinese markets. **

"We're going to invest like crazy in training and research," Macron said. It also added that France is well-positioned in the field of artificial intelligence because of its access to talent and the creation of startups around the technology.

Macron's words are true. With a "hope of the whole village", France already has the ambition to dare to challenge the global artificial intelligence.

Just in May of this year, a Paris-based company named Mistral AI, which had just been established for 4 weeks, completed a seed round of 105 million euros (about 827 million yuan) and was valued at 240 million euros (about 1.89 billion yuan). And the most powerful thing is: **This company has only 6 people including the founder until now, and only a 7-page PPT is used for financing, and even the finished product will not be launched until next year at the earliest! **

At the same time, the company's investment lineup focuses on luxury. In the first round of financing, 14 investors were squeezed in, including a group of top veteran VCs in the United States, VCs under the European "Old Money", as well as VCs from various European countries, as well as major well-known companies and executives.

Institutions scramble to fund Mistral AI

It has been established for a month, and real money will be paid to the account. Of course, this is not a novel or science fiction, but because the "golden fathers" have taken a fancy to the ambition of this company.

France and continental Europe hope to create a viable alternative to Silicon Valley AI companies such as Microsoft-backed OpenAI and Google's DeepMind. And it bears the hope of the whole of Europe. As Arthur Mensch, the company's chief executive, said, "There is growing recognition that this technology is transformative, and Europe needs to act on it as regulators, customers and investors."

As the founder and CEO, he met the other two founders, Timothée Lacroix and Guillaume Lample, during his studies at the Paris Polytechnic Institute and the Paris Higher Normal School. Although these three Frenchmen are young, only in their early 30s, they already have some experience in the AI industry.

These three young people can be said to be "battle-tested": Arthur Mensch was a researcher at DeepMind, an artificial intelligence company under Google, and a major contributor to the Retro, Flamingo and Chinchilla projects, gaining valuable experience in optimizing large language models. The other two founders were previously members of the Meta artificial intelligence team, and Lample even led the development of Meta's large language model LLaMA.

Such a team background is considered top in the field of AI. Because there are very few people in the world who really know how to construct the LLMs model of ChatGPT,** and they master the most difficult technology in the generative AI model, that is, the generative model itself. **At present, there are actually only a handful of people like them in the world who have the expertise to train and optimize large models.

Although ambitious, the training cost of artificial intelligence is increasing rapidly, and Mistral AI's products will not be launched until 2024. It is difficult to compete directly with OpenAI, which has a solid foundation. According to the financing memorandum, the company focuses on differentiated "dislocation competition". To sum up, there are 3 points of differentiation: **Open source, ToB and the European market. **The ultimate goal is to "construct safe, controllable and efficient technology, so that human beings can benefit from this scientific breakthrough".

Mistral AI's green light on financing all the way shows that France is full of confidence in leading the development of artificial intelligence in Europe. Of course, Britain is not to be outdone.

**UK: Loud thunder, little rain? **

In March, the British government pledged to spend 1 billion pounds ($1.3 billion) on supercomputing and artificial intelligence research, hoping to become a "tech superpower". As part of the strategy, the government said it wanted to spend around £900m building a "hyperscale" computer capable of building its own "BritGPT", to rival OpenAI's generative AI chatbot.

However, some officials have criticized the funding pledge as not doing enough to help the UK compete with giants such as the US and China. "It sounds great, but it's nowhere near where we need to be," Javid, a former government minister in Johnson's cabinet, said during a fireside discussion at London Tech Week.

While the UK doesn't appear to have a challenger like Mistral AI just yet, that doesn't mean the country doesn't have notable AI startups.

Digital media platform Synthesation, for example, lets users create AI-generated videos. According to foreign media reports, ** The company has received US$90 million in financing from investors including the US chip giant (Nvidia), which valued the company at US$1 billion. **

Are the Brits bad at building AI start-ups? The answer seems to be no, and the famous "unicorn" Inflection AI is an example.

Inflection AI, a startup led by British-born Mustafa Suleiman, a former DeepMind executive, has raised $1.3 billion in funding, backed by Microsoft, Nvidia and billionaires Reed Hoffman, Bill Gates and Eric Schmidt, among others.

The company's core product is a personal AI called Pi, a new category of AI designed to serve as an empathetic companion, offering conversation, friendly advice, and succinct information in a natural, fluid style. Unfortunately, the company is headquartered in California.

Combating Abuse of Artificial Intelligence

A big difference between the UK and France is how the two countries have chosen to regulate artificial intelligence and the existing laws affecting this rapidly developing technology.

For continental Europe, the European Union has enacted the Artificial Intelligence Act, which will be the first comprehensive law in a western country on the theme of artificial intelligence. In June 2023, lawmakers in the European Parliament approved the legislation.

The bill evaluates different applications of artificial intelligence based on risk. For example, real-time biometric and social scoring systems were deemed to pose "unacceptable risks" and were therefore banned in the regulations.

According to foreign media reports, Minesh Tanna, the global head of artificial intelligence at an international law firm called Simmons & Simmons, said that France will be directly governed by the law** but it would not be surprising if the relevant French regulators (whether it is the French National Information and Liberty Commission, abbreviated as CNIL, or the new AI-specific regulator) take an "aggressive approach" to implementing the law. **

According to the "Xinmin Weekly" report, on June 9 this year, Macron privately met with dignitaries such as the French Minister of Digital Technology, as well as AI experts from technology giants such as Facebook's parent company Meta and Google, planning to establish an AI regulatory agency in France. On May 16 this year, CNIL released an artificial intelligence action plan, which is divided into four aspects:

1. Understand the workings of AI systems and their impact on individuals;

2. Support and regulate the development of privacy-respecting artificial intelligence;

3. Integrate and support innovators in the French and European ecosystem;

4. Audit and monitor AI systems to protect individuals.

Corresponding to the European Union, in the UK, the government has not yet enacted laws specific to artificial intelligence, but has issued a white paper to make recommendations on how the various industry regulators should implement the existing rules of their respective industries. The white paper takes a principles-based approach to regulating AI.

The government has described the framework as a "flexible" approach to regulation, which Tanna believes** is more conducive to innovation than the French approach. **

In his understanding, the UK’s approach in a post-Brexit world is driven by a desire to encourage investment in AI, which gives the UK more freedom and flexibility to develop regulation at the appropriate level to encourage investment. Conversely, the EU’s AI Law could make France less attractive for AI investment because it creates a onerous regulatory regime for AI.

**Who will win? **

Anton Dahbura, co-director of the Johns Hopkins Institute for Assured Autonomy, said France definitely had a chance to lead Europe, but it faced stiff competition from Germany and Britain. Alexandre Lebrun, chief executive of Nabla, which provides artificial intelligence assistant services for doctors, said that the UK and France may be comparable in terms of the attractiveness of starting an artificial intelligence company. **

According to it, there is a good talent pool, strongholds such as Google and Meta artificial intelligence research center, and a reasonable local market, but it also warns that the EU AI law will make it "impossible" for start-ups to establish artificial intelligence in the EU. If the UK passes a smarter law at the same time, it will surely win the competition with the EU and France.

London, meanwhile, has been a doom and gloom for some areas of artificial intelligence:** For tech entrepreneurs, the UK is often unattractive. ** Opposition Labor leader Keir Starmer told attendees at London Tech Week that a series of crises in the country had generally dampened investor enthusiasm for technology. Investors are not investing in the UK because they do not see the conditions of political certainty required.

But Claire Trachet, chief financial officer of French tech start-up YesWeHack, said both the UK and France had the potential to challenge the dominance of US AI giants, but that would require both cooperation among European countries and competition between different centers.

According to her, to truly have a meaningful impact requires the European tech superpowers to come together, work together, leverage collective resources, collaborate with each other, and invest in fostering a strong ecosystem.

Trachet also added that combining the strengths of all parties, especially Germany's participation, can allow European countries to create a compelling alternative in the next 10-15 years, absolutely disrupting the artificial intelligence landscape, but this also requires a high degree of strategic vision and a collaborative approach.