Unleashing AI in Europe: Academia & Investors

October 10, 2022
We believe four main domains need to be adressed in order to pave the way for even more AI made in Europe (Part 1 of2)

This article was co-op project between ETH AI Center and Merantix; and co-written by Rasmus Rothe, Alex Ilic, Helga Rietz, Henry Schröder, Janette Wiget, Laurenz Lankes and Raphael Stücheli.

The global AI landscape is taking shape

The race for global AI leadership has quickly turned into a two-country game between the US and China:

In 2021, private investments in AI from the US amounted to almost $53 billion, in front of China, with $17 billion and Europe with $6 billion in funding. With 106 newly funded AI companies in 2021, Europe also lags behind the US and China who funded 299 and 119 new AI companies respectively. When it comes to AI patents filed, Europe only filed around four percent of globally filed AI patents in 2021 while the US filed 17 and China 52 percent.

Europe's dormancy in this battle for the technology of the future risks the continent losing touch if it does not act. The consequences of such a failure would have serious implications for the continent's competitiveness, political stability and prosperity. Although European Policy Makers have recently shown ambitious plans to actively take part in the race for AI leadership, very few experts see this happening in the near future.

The untapped potentials of Europe

However, not all hope is lost. Many ingredients to make Europe a successful location for and a booster of AI already exist. Europe has enormous potential for the development and application of AI. It has more AI researchers and developers per capita than both the US and China, and is home to many global industry leaders with immense domain expertise. Furthermore, it is developing a regulatory framework that is becoming the leading document for regulating AI. It is now in Europe's hands to use these advantages properly.

However, there are still big steps to take for Europe. Currently a very fragmented market, non-digitized companies, limited availability to the necessary data, unclear regulation, insufficient investments and lack of technical and hardware products prevent the success of AI in Europe.

Time for action is now

The vision of Europe catching up with the US and China in the race for the future of AI and acting as a pioneer in the long term requires immense and, above all, rapid response from all stakeholders involved; namely Academia, Investors, Corporations and Policy Makers. Therefore, we call for the following actions from these stakeholders to overcome these challenges:

1. Academia: Providing an ideal climate for innovation

  • Research vision aligned with European values:

Progress in AI is often measured in comparison with human performance. By shifting the course of AI research towards synergy rather than competition, we propose an artificial intelligence that enhances and collaborates with human intelligence. This mindset of Collaborative AI would act as a focal point for aligning research goals more closely with European values and act as a positive catalyst for a pragmatic approach to regulation while safeguarding strong individual rights.

  • Interdisciplinarity and joint European efforts accelerate break-throughs:

By bringing together researchers from AI foundations with researchers focused on AI applications, we will likely see many more innovations happening in shorter periods of time. A European AI network such as ELLIS can act as a lighthouse to connect the best minds and play to the strengths of each institution across Europe. It connects the local ecosystems around each institution and broadens the dialogue between academia, industry and start-ups at scale.

  • Enable a start-up friendly breeding ground:

AI is likely to create a variety of new jobs and business models that we cannot foresee yet. Typically, the majority of the new growth areas originate from innovative new start-ups. That is why  it is now more critical than ever that institutions ensure that gaining entrepreneurial experience is not just an optional extracurricular activity but takes up a key element aligned with all AI related study and research programs. Entrepreneurship should be also seen as a team effort, where programs that build collaborative, resilient, impact-oriented and diverse teams across boundaries of institutions will have the most profound impact.

2. Investors: A call for immense and rapid investments

  • Need for private investment:

The European Investment Bank calculates that investments of EUR 5-10bn annually are needed to close the investment gap to the US and China. While there are clear reasons for this underinvestment (high initial investment needs, low visibility of commercial applications, longer payback periods), the economic interests at stake are by no means insignificant and investors across the board must be incentivized to actively participate in closing this investment gap. There are ample ways how private investments can be encouraged: Tax benefits for angel investors to spur participation, less punishing bankruptcy regimes to decrease the cost of failure, and clear signs of commitment that governments are doubling down on the technology and its future to give investors peace of mind.

  • Need for public investment:

Due to the long payback periods associated with investments in AI infrastructure, such as investments in high performance computing clusters or computer chips, the private markets have little incentive to finance the necessary infrastructure. Henceforth, states have to lay the groundwork. Authorities need to scrap their bureaucratic structures and become entrepreneurial themselves to seize the long-term value of AI - similarly to how the GPS, microchips, touch screens, and the internet were initially funded projects by the US government. The format in which authorities invest in AI can take different shapes. Exemplary options could be to develop AI solutions and infrastructure “in-house”, or to provide venture funds with more capital to deploy. However, it is crucial that these public investments are being allocated towards high-risk, high-reward projects that otherwise would not have received funding in the private market.

  • Governments as clients:

Instead of relying on (public) funding, a startup can benefit a lot more by showing strong revenues. This is where governments can prove to be a huge lever. By becoming clients of new ventures, the government on one side gets quick access to digital solutions, while startups can prove strong early traction and gain insights to finetune their product. Subsequently, that traction proves very helpful in attracting foreign investments in later funding rounds.

In the second part, we will go into the domains "Corporations" and "Policy Makers", and give an outlook on the current status.

Read more