4
Executive summary
ments (e. g. GDPR) have hindered possibilities
of industrial data sharing. The EU also struggles
to develop and retain key data science talent.
Although European institutions produce world-
class talent and research in AI-related elds, they
have yet to reach the scale or inuence of US and
Chinese institutions, and much of the talent they
develop has migrated to those two countries. Nor
does the EU possess a deep reserve of high-end
computing power, a fundamental requirement for
world-class AI innovation at scale. Finally, while
the climate for commercialization varies from one
EU member state to the next, the overall ecosys-
tem for innovative risk-taking, technology trans-
fer, venture investment and startup growth lags
behind that of global AI leaders.
Nevertheless, many strengths remain, and they
underpin the EU’s continued emergence as a crit-
ical player in the science, geopolitics and ethics
of AI and related elds. To the extent it coalesces
and becomes available to developers, its com-
mon market can generate a deep pool of data for
cutting-edge R&D. Its leading research institutions
still develop world-class AI talent, and the increas-
ing digitalization of the existing industrial power
base is starting to generate more local opportu-
nities for those experts. Furthermore, the region
continues to lead the world in its awareness of
and emphasis on human-centric, private and eth-
ical uses of AI and data science. These are critical,
indispensable strengths on which the EU – and, in
many respects, the world as a whole – will rely in
the decades to come.
However, these advantages are not enough to
enable the EU to stand on its own as a “Third
Way” alternative to the US and China. Ultimately,
countries will have to individually or collectively
align, at least in part, with a US or Chinese mindset
regarding technology, geopolitics, and economic
development. We have argued elsewhere that the
EU best aligns with the liberal democratic ideas
embodied in the US constitution. For the purposes
of this report, however, we have focused on the
The European Union (EU) and its members have
recognized the potential for articial intelligence
(AI) to drive economic, business and societal pros-
perity. Critically, they have also recognized many
of the risks that accompany AI and the various
applications and systems it empowers. Many of
these considerations are reected in the various
national and EU-wide AI strategies and standards.
Perhaps more than any other region or country
in the world, Europe has made human rights and
privacy the “North Star” of its strategies, part-
nerships, governance, and commercialization of
advanced technologies.
This has become a primary strength as the EU
and its members develop their AI ecosystems,
but it also drives many of the region’s key weak-
nesses. Perhaps nothing exemplies this duality
better than the General Data Protection Regula-
tion (GDPR). While the GDPR has become a global
standard for the preservation of individual data
privacy and a key check on the hegemonic power
of the large digital service platforms, its structure
has also curbed innovation, commercialization,
and the collection of massive data pools that drive
the development and training of AI systems. Care-
ful consideration of ways to calibrate and recali-
brate their approach to partnerships, governance
and commercialization will allow the EU and its
member states to expand their inuence on global
AI development, while fostering a domestic envi-
ronment that allows their companies and research
institutions to compete more eectively with the
United States of America (USA) and China.
Such calibrations must be based on a deliber-
ate and clear-eyed understanding of the factors
that currently limit AI development across the
EU. While the EU is home to 446 million resi-
dents – representing the third-largest market in
the world after India and China – a collective pool
of usable data has not yet coalesced to power AI
research and development (R&D). This is particu-
larly true for European industry, where concerns
about trade secrets and governance require-
