Solvency deficits – The EU is fragmented and cannot harmonize markets; EU action creates Brain Drain to the US and UK; and their regulatory system slows implementation
Lawrence and Cordey, 2020 – researchers for The Cyber Project at the Belfer Center for Science and International Affairs [Christie and Sean, August, The Cyber Project Paper “The Case for Increased Transatlantic Cooperation on Artificial Intelligence Edited by Lauren Zabierek and Julia Voo https://www.belfercenter.org/sites/default/files/2020-08/TransatlanticAI.pdf Acc. 4/21/22 TA]
Challenges Due to the socio-technical transformative power of AI, governing and harnessing its benefits is a complex affair. There are three fundamental issues the EU needs to overcome: A fragmented market and landscape: This patchwork landscape results from the EU’s political configuration of 27 different Member States, all of which have different levels of maturity and competitiveness with regard to its AI ecosystem. Some, like France or Germany, are major players in their own right, but the majority of EU Member States carry little weight on their own.248 This disparity in capability drives the EU to pool its resources together. The UK’s decision to leave the EU constituted a serious blow to the EU’s standing, as it deprived it of one of its most mature, rich and innovative AI environments.249 In addition, European AI assets (e.g., talent, education, expertise, research, start-ups, and capital) are highly fragmented and decentralized.250 As Professor of Theoretical Philosophy Thomas Metzinger pointed out, this makes effective coordination of all the different stakeholders complex but highly important. One example where coordination could be improved is research funding both within the existing EU research frameworks and between the EU government, academia, industry, and the Member States.251 Europe is also fragmented along market and cultural lines. Indeed, compared to the US or China, the EU’s AI market is not yet consolidated nor harmonized. In addition, some major sectors of the economy such as automotive, agriculture, energy, as well as the public sector, present varying degrees of AI maturity, penetration, and integration.252 According to a McKinsey study,253 in 2017 only 25% of EU large enterprises and 10% of small and medium enterprises used big data analytics.254 This slow uptake of AI relates, in part, to the lack of trust of the general public and companies around issues of algorithmic transparency and biases. Meanwhile, the EU’s AI governance must also be able to navigate and transcend potentially divisive differences that come from the cultural, historical, strategic, and institutional differences that characterize each and every Member State.255 Underlying structural factors impede its development and competitiveness: Europe lacks tech giants that characterize the US and Chinese tech landscape. As a result, European start-ups and tech companies compete against peers that have not only considerably larger investment capabilities, which enable them to acquire the latest technologies and companies, but also provide ample access to data and a greater ability to attract and retain a skilled workforce.256 The lack of easily available data, which impedes innovation, has emerged as a key issue for the EU and its private sector companies.257 These concerns are additionally fueled by the relatively more restrictive European privacy laws—and maybe soon mandatory ethical guidelines.258 “Brain drain,” meanwhile, poses a unique challenge to the EU as promising European researchers often choose to move to the US, Canada, or the UK for academic opportunities. Regarding the former for instance, 19% of Europe’s undergraduates move to the US to study, while 14% of European graduates move to the US to work. Overall, 11% of the US’s top tier AI workforce—which represent 59% of the global workforce— comes from Europe.259 In some cases, AI talent also moves to the US to work at larger international companies that acquire their start-ups.260 Examples include the French Moodstock (acquired by Google) and the UK’s Magic Pony Technology (acquired by Twitter).261 A strong but slow regulatory process: While the policy process regarding AI has, under the impetus of the EC, accelerated over the last few years, the overall regulatory process cannot be as easily accelerated and might take years. According to the EC’s calendar (see Figure 8), the AI-related regulatory process will begin at the end of this year (2020), but the actual drafting and passage of the associated legislation is likely to take several years. As such, the EU runs the risk of not keeping up with the pace of technological evolution in AI as well as further politicization of regulation similar to the process of passing GDPR. Finally, as with most democracies, the EU will also face the challenge of operationalizing related regulations and principles, carefully and skillfully, balancing private and public sector interests throughout the process.