In the European Union, a lot of political declarations have been made about the need to bolster AI research and startup scenes. However, these statements almost always come with a “but.”
Take, for example, Thorbjørn Jagland, Secretary General of the Council of Europe. His statements from the 2019 Helsinki Conference on Artificial Intelligence serve as a kind of archetype:
Artificial intelligence will revolutionize the way we live. In fields such as medicine, communications, and transport new opportunities abound. But the consequences of AI’s advance for democracy, human rights, and the rule of law are still unclear.” -Thorbjørn Jagland (emphasis added)
Jagland’s statements typify how the EU feels about AI. There may be opportunities, but these opportunities are quickly overshadowed by risks. Europe wants AI, but then again, it isn’t sure it wants AI. In the end, the EU is overwhelmed by the ethical implications of a booming AI market.
The ramifications of Europe’s tendency towards an overly cautious approach to AI is already clear. In an interview with Sifted earlier this year, Kai-fu Lee, a former Apple and Google executive, stated that he doesn’t even think the EU will be able to win a “bronze medal” when it comes to AI. “European artificial intelligence is losing the race,” he said.
However, here’s the thing — Europe needs a robust AI ecosystem. A 2018 study by McKinsey Global Institute found that AI alone could boost Europe’s annual GDP growth by 1.2 percent through 2028. And a Tractica report from 2017 declared that, through 2025, worldwide revenue from the AI market would grow 25-fold. Yet, if Lee’s fears are realized and Europe fails to place in the AI race, the sociopolitical consequences will be far more troubling than any economic repercussions, as the EU will find itself forever subordinated to (and dependent on) foreign nations’ algorithms.
Through the General Data Protection Regulation (GDPR) and similar regulatory frameworks, Europe has attempted to stymie the unchecked influence enjoyed by tech behemoths in China and the US. But this will not be enough. If Europe is to succeed, it needs its own innovation ecosystem and to author its own algorithms.
Fortunately, time is not yet out. Here are some of the most pressing hurdles and how they can be overcome.
Understanding the EU’s AI Woes
Statistics that examine technological development and proficiency across Europe tell a story of legacy systems and varying degrees of openness to new tech. More troubling, these analyses reveal that an alarming number of Europeans can’t productively contribute to tech-based industries. For example, although digitalization will affect all areas of the labor market, a staggering 44% of European citizens do not have basic digital skills, versus just 14% in the US.
Although Europe possesses a strong research tradition that is sustained by an energetic spirit of inquiry, it seems that this spirit becomes fatigued and disinterested when it comes to AI. Notably, this statement applies to individuals and universities equally. Figures from the World Intellectual Property Organization show that there has been a surge in AI patent applications globally, with annual filings tripling from 2014 to 2019. However, this research was dominated by institutions in China, the US, and South Korea. Among the top 167 universities and public research institutions for patents, Europe houses only four.
When it comes to AI-based companies, progress is just as stilted. A Barclay-backed MMC Ventures report from 2019, which analyzed over 2,800 AI startups in Europe, revealed that two in five of these demonstrate little evidence of actually using AI in their products and services. It also found that just one in every twelve newly founded tech companies is an AI company.
Then there are financial issues. In China, the government is planning to invest more than $30 billion in AI and related technologies through its latest venture capital fund — and this is just through its latest fund. In the US, just six tech companies invested more than $54 billion in AI research in 2015 alone. To say that Europe lags behind is to utter a dramatic understatement. The constituent countries are hoping to achieve $22 billion in public and private investments in AI research by the end of 2020.
Although the European Commission plans to invest an additional $1.6 billion through its Horizon 2020 program, and in an effort to turn France into a “startup nation” President Emmanuel Macron announced another $1.6 billion in public funding for artificial intelligence by 2022, these efforts fall short of US and Chinese volumes.
Adding to all of this is the fact that, even if entrepreneurs, policymakers, research organizations, and other institutions come together in a way that is perfectly timed and optimized for efficiency and effectiveness, still, the success would be difficult and hard-won. This is because the US and China both have ecosystems where AI technologies can forage for synergies that quickly lead to exponential growth. In these regions, there have already been several decades of interdisciplinary collaboration. Semiconductor companies help PC developers, who help software engineers, who help internet and social media engineers, who help mobile companies, and on and on. And now, all these resources are helping AI startups.
The way forward
Yet, there’s a great case for the EU to rise to the top in the race for AI dominance. With its rich history of research rigor, and its unique positioning on data privacy, increased dialogue between engineers and policymakers could catapult the continent to the forefront in a matter of years. With its idealism, seasoned with a spirit of pragmatism, and a focus on innovation, Europe possesses, as it did in previous industrial revolutions, the potential to be pioneers.
Here’s what needs to be done to secure this success.
- The first step is for Europe to endow its citizens with basic digital skills. It goes without saying that the baseline for even competing in the AI race is a technologically literate population. This education needs to begin at the elementary level and continue through graduation, which means establishing working groups to evaluate and, where and when needed, iterate new computer science programs.
- Along these same lines, the EU must work to ignite a spirit of inquiry and a persistent sense of curiosity when it comes to AI, particularly when it comes to universities and research programs.
- The next step is to offer incentives for individuals who are creating AI-based companies and those who are otherwise incorporating AI into their fundamental frameworks. This logically leads to the next point, which is to…
- Create government-backed heavy incentivizing for the private funding of AI-specific startups in order to attract private equity into the playing field, and allow research to bloom under policy protections.
- Experiment in the underdeveloped European markets where there is practically no legacy competition. For example, the more rural regions of Europe or the former Soviet satellite regions.
- Use an existing infrastructure that people love or need to use, and build something on top of it that is totally different. In this case, Uber and Airbnb serve as ripe examples. They took something that people love and need (transport, travel, and accommodations) and imagined new ways that these forms of infrastructure could be optimized and used. By building something totally out of the box, these companies were able to create a whole new market category and develop it to dominance.
- Utilize European knack for standardization to create an AI standard. Europe could see the current worldwide lack of standardization as an opportunity and participate with full force in its creation.
- GDPR is a unique legislative development that no societies have been able to create to date. Leverage it to form European-branded AI with a market reach outside of the continent. Europe can use its idealism to find the next paradigm shift and apply its innovation there.