El enfoque europeo para construir la política y la gobernanza de la IA: ¿un refugio para burócratas o innovadores?

AutorOlga Shumilo, Tanel Kerikmäe
CargoUniversidad Tecnológica de Tallin (TalTech)/Universidad Tecnológica de Tallin (TalTech)
Páginas1-14
https://idp.uoc.edu
IDP No. 34 (December, 2021) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
1
2021, Olga Shumilo and Tanel Kerikmäe
of this edition: 2021, Universitat Oberta de Catalunya
Dossier “Europe facing the digital challenge: obstacles and solutions”
The European approach to building
AI policy and governance: a haven
for bureaucrats or innovators?
Olga Shumilo
Tallinn University of Technology (TalTech)
Tanel Kerikmäe
Tallinn University of Technology (TalTech)
Date of submission: May 2021
Accepted in: November 2021
Published in: December 2021
Abstract
Disruptive technologies and the domination of digital platforms have challenged the global economy
players twice — f‌irst, to get a hand on them, then to mitigate the possible risks. It is beyond doubt
that reliable artif‌icial intelligence (AI) can bring many benef‌its at the European level, such as better
health care, safer and cleaner transport, more eff‌icient manufacturing, and sustainable energy. But
regulating the unknown requires considerable effort on how to attract investors using clear rules while
keeping human control over the algorithms as a priority. In April 2021, the EU Commission published
a holistic proposal to regulate the use of AI, which promises to put trust f‌irst and ensure that facial
recognition and big data operators will never abuse fundamental human rights. Although the proposal
is likely to be amended during EU-wide discussions, the new approach to AI will clearly give citizens the
reassurance to adopt these technologies while encouraging companies to develop them. Hence, this
article aims to map the core challenges for the EU policy on the use of AI, as well as the milestones of
developing the holistic legislative proposal, and clarify if the afore-mentioned proposal indeed solves
all the AI-related risks for future generations.
Keywords
AI; artif‌icial intelligence; EU law; fundamental rights
https://idp.uoc.edu
Universitat Oberta de Catalunya
The European approach to building AI policy and governance: a haven for bureaucrats or innovators?
IDP No. 34 (December, 2021) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
2
2021, Olga Shumilo and Tanel Kerikmäe
of this edition: 2021, Universitat Oberta de Catalunya
El enfoque europeo para construir la política y la gobernanza
de la IA: ¿un refugio para burócratas o innovadores?
Resumen
Las tecnologías disruptivas y el dominio de las plataformas digitales han desaf‌iado a los actores de la
economía global dos veces: primero, para echarles una mano, luego para mitigar los posibles riesgos.
Está fuera de toda duda que la inteligencia artif‌icial (IA) f‌iable puede aportar muchos benef‌icios a nivel
europeo, como una mejor atención sanitaria, un transporte más seguro y limpio, una fabricación más
ef‌iciente y una energía sostenible. Pero regular lo desconocido requiere esfuerzos considerables sobre
cómo atraer inversores con reglas claras y al mismo tiempo mantener el control humano como una
prioridad sobre los algoritmos. En abril de 2021, la Comisión de la UE publicó una propuesta holística
para regular el uso de la IA, que promete poner la conf‌ianza en primer lugar y garantizar que los ope-
radores de reconocimiento facial y big data nunca incumplan los derechos humanos fundamentales.
Aunque es probable que la propuesta se modif‌ique durante los debates a escala de la UE, el nuevo
enfoque de la IA dará claramente a los ciudadanos la conf‌ianza para adoptar estas tecnologías, al tiem-
po que animará a las empresas a desarrollarlas. Por lo tanto, este artículo tiene como objetivo trazar
los principales desafíos para la política de la UE sobre el uso de la IA, así como los hitos del desarrollo
de la propuesta holística legislativa, y aclarar si dicha propuesta resuelve realmente todos los riesgos
relacionados con la IA para las generaciones futuras.
Palabras clave
IA; inteligencia artif‌icial; derecho de la UE; derechos fundamentales
https://idp.uoc.edu
Universitat Oberta de Catalunya
The European approach to building AI policy and governance: a haven for bureaucrats or innovators?
IDP No. 34 (December, 2021) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
3
2021, Olga Shumilo and Tanel Kerikmäe
of this edition: 2021, Universitat Oberta de Catalunya
Introduction
Since their emergence, artif‌icial intelligence (hereinafter —
AI) systems have become a topic of hot debate, fuelled by
a lack of understanding of algorithms and their potential
to change the world for the better. In 1973, the so-called
Lighthill report buried the boom of AI research in Great
Britain for a decade, based on the rapporteur’s observa-
tion that building able robots has so far failed to deliver
any practical value and thus had to be discontinued for the
sake of funding more relevant research (Agar, 2020). The
21st century has brought great success to the developers
of automated solutions, and digital platforms are more
than eager to employ disruptive technologies to skyrocket
operational eff‌iciency (Arrieta et al., 2020). The AI has also
challenged us to redef‌ine the phenomena of non-discrimi-
nation and fairness (BBC, 2021) since the proverbial errare
humanum est formula reminds us of the eternal prejudice
behind public administrations. The rapid adoption of the
technology in the public sector has led to inevitable mis-
takes that proved that the risks of predictions provided a
black box fed by human input. Apart from the notoriously
dystopian results of the AI-backed COMPAS system (Black
& Murray, 2019), we have to acknowledge the fact that
algorithms are powerful enough to transform any sphere
of human activity into a race for higher eff‌iciency, be it
anti-plagiarism plugins (Büttner et al., 2015) or weaponry
(Cabral, 2020). These and other risks have triggered the
discussion on how to elaborate a human-centred approach
to developing these technologies, which coined terms like
XAI (eXplainable AI) (Arrieta et al., 2020) and responsible
AI (CEN-CENELEC, 2020). Naturally, lawmakers across the
globe were eager to pioneer a solution that would please
all the stakeholders. Still, it was of particular signif‌icance
for the EU, given the high priority of human rights protec-
tion enshrined in the Charter (Charter of the Fundamental
Rights of the Union, 2012) and the worldwide recognition
of the mechanisms established by the General Data Pro-
tection Regulation (2018).
Considering the above, the current article aims to portray
the institutional approach for def‌ining AI, as well as the
driving forces and barriers behind the EU AI policy in
1. EUR-Lex-12012P/TXT. Charter of Fundamental Rights of the European Union. Official Journal C326, 26 October, pp. 391-407.
2. EUR-Lex-32001L0095. Directive 2001/95/EC of the European Parliament and of the Council of 3 December 2001 on General Product
Safety, Official Journal L 11, 15 January 2002, pp. 4–17.
chapter 1, in order to offer a critical perspective on the AI
policy steps taken by the main EU stakeholders so far, with
a specif‌ic focus on the recent horizontal legislative pro-
posal for regulating AI, in chapters 2 and 3. Consequently,
we raise the research question on whether the new Euro-
pean regulatory framework is answering the technology
challenges and able to foster international cooperation
in the f‌ield. Methodologically, we rely on desktop-based
literature analysis and assessments by innovation experts
to substantiate our evaluation of the EU AI policy concern-
ing the mentioned aims. The authors strive to bridge the
existing gap in the literature on the forthcoming AI regula-
tion proposed by the European Commission and scrutinize
any possible drawbacks that the draft may have.
1. Deningthedriversandbarriers
of EU-wide AI systems regulation
policy
Modern literature offers a plethora of def‌initions for AI,
or ‘thinking machines’, since back in Alan Turing’s era it
was hard to come up with a unanimous understanding
for words such as machin’ and to think.
1
For example,
the modern computer science doctrine offers a working
def‌inition for AI as a ‘similarity between a typical human
mind H… and a typical intelligent computer C…, in which
the two are described with a certain level of abstraction’
(De Hert et al., 2016). But legal interpretations ought to
be inclusive, precise, comprehensive, and practical (Del-
ponte, 2018), i.e., to unequivocally delineate the threshold
degree of a particular agent’s autonomous ability to mim-
ic human behaviour. This is also why policymakers focus
mainly on so-called ‘weak’ AI — the development of a fully
autonomous AI or artif‌icial general intelligence (AGI) is
theoretically challenged by many,
2
or at least not foreseen
to arrive anytime soon. However, such a creature would
def‌initely require a new layer of legislation covering its
function in human society.
The def‌inition of AI in EU strategic documents also varies,
from “a collection of technologies that combine data, algo-
https://idp.uoc.edu
Universitat Oberta de Catalunya
The European approach to building AI policy and governance: a haven for bureaucrats or innovators?
IDP No. 34 (December, 2021) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
4
2021, Olga Shumilo and Tanel Kerikmäe
of this edition: 2021, Universitat Oberta de Catalunya
rithms and computing power” to “software that is devel-
oped with one or more of the approaches and techniques
listed in Annex I and that, for a given set of human-def‌ined
objectives, can generate outputs such as content, predic-
tions, recommendations, or decisions affecting the envi-
ronments they interact in” (AI Regulation Proposal, 2021).
The latter is complemented with the following Annexes
that help to def‌ine the taxonomy of AI technologies:
Machine learning (including supervised, unsupervised,
and reinforcement learning, using a wide variety of
methods including deep learning);
Logic- and knowledge-based approaches (including
knowledge representation, inductive (logic) program-
ming, knowledge bases, inference/deductive engines,
(symbolic) reasoning, and expert systems);
Statistical approaches, Bayesian estimation, search
and optimization methods.
As we can see from the list above, the legislators are
trying to avoid any possible loopholes and even classify
‘statistical approaches’ as AI, although few will attribute
machine learning features to regression analysis, which
provides the simplest data-based predictions for ‘tradi-
tional statistics’. On the other hand, the concept fails to
cover the technologies of quantum computing or ‘edge’
computing.
3
Furthermore, the proposal includes a provi-
sion for the ‘regulatory sandboxes’ that supposedly would
not hinder the development of innovations (art. 54), which
albeit raises questions on possible data protection loop-
holes, as it is not clear if the sandbox eliminates the GDPR
mandate over the experiment. It is also hard to disagree
with the meticulous analysis performed by S. Ranchordas
(2021) regarding the dependence of sandbox success on
the quality of the experiment design and its compatibility
3. EUR-Lex-32006L0042. Directive 2006/42/EC of the European Parliament and the Council of 17 May 2006 on machinery and amending
Directive 95/16/EC (recast), Official Journal L 157, 9 June 2006, pp. 24–86.
4. EUR-Lex-32014L0053. Directive 2014/53/EU of the European Parliament and of the Council of 16 April 2014 on the harmonization of the
laws of the Member States relating to the making available on the market of radio equipment and repealing Directive 1999/5/EC Official
Journal L 153, 22 May 2014, pp. 62–106.
5. EUR-Lex-52018DC0237. European Commission. 2018, April 25. Communication from the Commission to the European Parliament, the
European Council, the Council, the European Economic and Social Committee and the Committee of the Regions Artificial Intelligence
for Europe [online]. Available at: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=COM%3A2018%3A237%3AFIN
6. European Commission (2020). “White Paper on Artificial Intelligence: a European approach to excellence and trust”. In: European
Commission [online]. Available at: https://ec.europa.eu/info/publications/white-paper-artificial-intelligence-european-approach-excellen-
ce-and-trust_en
with the rule of law principles — fair law ceases to exist
once it becomes incomprehensible.
Potential caveats and threats of AI were made at the
dawn of its development and inspired Isaac Asimov to
formulate the Three Laws of Robotics.
4
Although the
spectrum of AI opportunities, risks, and dangers has
been extensively analyzed in recent years (Dwivedi,
2019),
5
we would like to highlight the specif‌ic factors
that have been pushing forward the AI agenda at the
EU level. Firstly, the EU f‌inds itself in a race with the
two other superpowers (the US and China) to domi-
nate in attracting innovation while remaining a legal
trendsetter. Secondly, the bloc is evidently eager to
avoid a patchwork of national laws on AI that would
undermine the global competitive ability of the Union.
Minor discrepancies have already been showcased in
the national AI Strategies of Member States, although
most of them declared that they intended to follow a
multi-lateral European approach in the future,
6
if it is
developed. Thirdly, as AI systems are often reliant on
Big Data, they are already addressed to some extent by
the GDPR (2018), but def‌initely require a more detailed
regulation on such issues as those related to the deploy-
ment of the technology in law enforcement, healthcare,
employment, education, public services, trade, and
critical infrastructure to prevent potential harm. Lastly,
the rapidly developing f‌ield needs a timely legislative
approach, as research often progresses faster than the
bureaucratic mechanisms of approving a multi-lateral
agreement. These circumstances largely explain the
urge of the European Commission (hereinafter — EC)
to present a proposal as soon as April 2021, only three
years after issuing the f‌irst communication on this topic
(in comparison, the developing of GDPR started in 2011).
https://idp.uoc.edu
Universitat Oberta de Catalunya
The European approach to building AI policy and governance: a haven for bureaucrats or innovators?
IDP No. 34 (December, 2021) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
5
2021, Olga Shumilo and Tanel Kerikmäe
of this edition: 2021, Universitat Oberta de Catalunya
2. The European policy response
to AI challenges (2018-2021)
As mentioned above, by early 2018 the EU found itself
lagging behind its competitors in terms of unanimous pol-
icy response to the rapid growth of AI-based innovations.
By then, the US government had already presented two
policy documents-National Artif‌icial Intelligence Research
and Development Strategic Plan
7
and ‘Preparing for the
Future of Artif‌icial Intelligence’ report,
8
with a wide array
of planned research, especially in the area of military
technologies (Evans, 2020). Being one of the leaders in
AI research, with its prominent universities and start-ups,
accounting for 32% of the global AI research market (In-
trona, 2016), the EU fell short of outperforming the US
and China in private investment in the sector.
9
Thus, the
policy action at that point was caught between two op-
posing approaches offered by the ‘ethicists’, or stringent
GDPR proponents on the one hand, and ‘self-regulation’
advocates on the other hand.
10
As revealed later on, the
latter has been represented in EU Commission expert
circles to a greater extent (Johnson, 2019), although the
chances of making a U-turn towards informal regulation
have been largely decreased by the enacted GDPR and its
core concept of the right to know why exactly a certain
data-based decision has been made. Further steps taken
by the European stakeholders can be organized into the
following streams — communication of the strategic plans
and debating of the proposals, research on risks and ben-
ef‌its, and legislating.
1.1. Communication
Debates on the need to regulate the f‌ield of AI at supra-
national level have off‌icially kicked off with the arrival of
European Parliament resolution with recommendations to
the Commission on Civil Law Rules on Robotics (2017), that
has set out the key principles concerning the development
of robotics and AI for civil use, addressed main spheres of
7. European Commission (2021). “Annexes to the Communication from the Commission to the European Parliament, the European Council,
the Council, the European Economic and Social Committee and the Committee of the Regions; Fostering a European approach to Artificial
Intelligence”. In: European Commission DS [online]: Available at: https://digital-strategy.ec.europa.eu/en/library/coordinated-plan-artifi-
cial-intelligence-2021-review
8. European Union Agency for Cybersecurity (2020). “Artificial Intelligence Cybersecurity Challenges”. In: ENISA [online]. Available at:
https://www.enisa.europa.eu/publications/artificial-intelligence-cybersecurity-challenges
9. ISO/IEC JTC 1/SC 42 Artificial intelligence [online] [2021]. Available at: https://www.iso.org/committee/6794475.html
10. ISO/IEC JTC 1/SC 42 Artificial intelligence [online] [2021]. Available at: https://www.iso.org/committee/6794475.html
its application, liability issues and urged the EC to report
to the European Parliament on the ‘latest developments
in robotics… on an annual basis’. Moreover, the document
has stressed the utmost importance of safeguarding hu-
man control over AI and resolving possible ethical issues,
e.g. an emotional connection between humans and robots.
This wake-up call has resulted in the EC issuing an ‘AI for
Europe’ Communication, that has marked the launch of
an ambitious European AI initiative, aimed at boosting
innovation in public and private sectors, along with cre-
ating ‘an appropriate ethical and legal framework’ (Kelly,
2020). The text of the EC Communication (and the original
Motion) offered to the reader a plethora of references to
images of artif‌icial intelligent creatures from mythology
and classic literature, clearly a pedagogical maneuver that
has however added confusion to the envisioned AI-driven
future (Lee, 2018). The ambitious document paved the
way towards creating an inclusive European AI Alliance
discussion forum, as well as a joint Coordinated Plan on AI
of the EC and Member States. Twenty-f‌ive Member States
later signed a Declaration of Cooperation on AI. Shortly
afterwards, the EC Joint Research Centre (JRC) together
with the Directorate General for Communications Net-
works, Content and Technology launched AI Watch — the
research monitor of policy initiatives and AI progress for
implementing the European Strategy for AI.
In April 2021, the Coordination Plan has been updated
with regards to the potential of AI ecosystems, at the
service of the Green Deal and to address the aftermath
of the COVID-19 pandemic, and the key priorities for the
future are: creating favourable conditions for putting the
locally developed AI products on the global market, secur-
ing fundamental rights protection and contributing joint
efforts to six vital policy areas — environment, healthcare,
robotics, law enforcement, infrastructure and agriculture
(Maas, 2019). It is worth noting that policy-making in the
EU has been recently subject to a paradigm shift towards
https://idp.uoc.edu
Universitat Oberta de Catalunya
The European approach to building AI policy and governance: a haven for bureaucrats or innovators?
IDP No. 34 (December, 2021) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
6
2021, Olga Shumilo and Tanel Kerikmäe
of this edition: 2021, Universitat Oberta de Catalunya
professionalization (Maedche et al., 2019) and, e.g., the
aforementioned policy areas selection was based on sug-
gestions from the Member States and practical evidence.
In 2019, the EC issued another Communication, titled
‘Building Trust in Human Centric AI’, reassuring the Un-
ion to place its trust on developing further guidelines
for trustworthy AI and encouraging stakeholders to give
feedback on the respective guidelines (McCauley, 2007).
Research
As mandated by the Communication on AI for Europe, the
EC has been annually reporting on the safety and liability
legal frameworks for AI, Internet of Things (IoT) and other
data-driven technical solutions, in order to identify the
gaps that will be further bridged. In 2020, the research
has drawn the attention of policymakers to the shortcom-
ings in the General Product Safety Directive, Machinery
Directive, the Radio-Equipment Directive and mentioned
the forthcoming ‘New Legislative Framework’. It has also
been indicated that national laws can potentially cover
the burden of proof issues in case of damage done by an
automated vehicle (Meyer, 2018).
Anyhow, the core part of the research on formulating AI
policy has been delegated by the EC to the newly estab-
lished High-Level Expert Group on Artif‌icial Intelligence
(hereinafter — HLEG), tasked to compose AI Ethics Guide-
lines and Policy and Investment Recommendations, that
will become a roadmap for further legislating process and
policy-making. Following the records of the group meet-
ings, although the group has been challenged because of
the experts diverse backgrounds and short deadlines, they
have managed to present an ambitious yet practical set of
rules and safe-guards and given a call to elaborate a ho-
listic European legal instrument (Nífhaoláin et al., 2020).
Another EU task force focused on the future of AI is the
Ad-Hoc Working Group on Artif‌icial Intelligence Cyber-
security at the EU Agency for Cybersecurity (ENISA),
which recently issued its f‌irst report on AI Cybersecurity
Challenges (Perrault et al., 2019). The document has de-
veloped a taxonomy of AI threats and allocated them into
the following blocks: nefarious activity, eavesdropping,
physical attacks, unintentional damage, failures, outages,
11. Abbr. for service-level agreement
disaster, and legal issues. The latter has particularly men-
tioned ‘corruption of data indexes, prof‌iling of end-users,
vendor lock-in, weak requirements analysis, lack of data
governance policies, lack of data protection compliance of
third parties, SLA
11
breach’ (Perrault et al., 2019).
Although the Council of Europe is a non-EU institution, the
overview of AI-related research and reporting activities in
Europe would not be complete without the recent Feasibil-
ity Study of its Ad Hoc Committee on Artif‌icial Intelligence
(CAHAI), issued in December 2020. The report has iden-
tif‌ied the challenges of AI systems for human rights, and
identif‌ied several possible tracks to address them:
1) To amend the existing binding agreements;
2) To adopt a new binding legal instrument
3) To develop soft law at the level of CoE or Member
States.
Regulation
Against the solid research background, in April 2021 the
EC has presented the long-awaited pioneer proposal for
a comprehensive EU AI regulation. As explained in the
memorandum, the goal of the Union has remained the
same — to build an innovator-friendly system of AI devel-
opment guidelines and rule out the risks of privacy abuse
and drooping into a dystopian techno-totalitarianism. The
scope of the proposal is relatively wide — both private
actors’ and public authorities’ activities, e.g., recidivism
likelihood scoring and grading school entrance exams will
be subject to thorough control. It is also remarkable that
the Commission does not rely only on the legislative re-
strictions to be imposed, but also encourages service pro-
viders to introduce their soft-law tools — codes of conduct.
These documents can be complemented by the internal
self-reporting and raising awareness measures taken by
the industry, which is also positive from the perspective of
the funding required to implement the proposal.
Special focus has been made on the proposed taxonomy
of AI systems — into those posing ‘minimal risk’, ‘limited
risk’, ‘high risk’ and ‘unacceptable risk’. The latter implies
violating fundamental human rights, e.g., governmental
scoring systems, live remote biometrical identif‌ication
(unless required by urgent emergency measures). If an
https://idp.uoc.edu
Universitat Oberta de Catalunya
The European approach to building AI policy and governance: a haven for bureaucrats or innovators?
IDP No. 34 (December, 2021) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
7
2021, Olga Shumilo and Tanel Kerikmäe
of this edition: 2021, Universitat Oberta de Catalunya
algorithm is ‘high-risk’, i.e. its use impacts fundamental
rights adversely, the authorizer is obliged to maintain:
Adequate risk assessment and mitigation systems.
High quality of the datasets feeding the system to
minimise risks and discriminatory outcomes.
Logging of activity to ensure traceability of results.
Detailed documentation providing all information nec-
essary on the system and its purpose for authorities to
assess its compliance.
Clear and adequate information to the user.
Appropriate human oversight measures to minimise
risk.
High level of robustness, security, and accuracy.
12
Limited risk is associated with cases in which AI can be
confused with a live person (chat bot), thus automated
tools must be revealed. Finally, the rest of AI systems can
be deployed without any restrictions. Future standard-
ization and implementation of the new guidance will be
delegated to a new watchdog institute, European AI Board.
The chosen risk-based approach to the regulation is
probably the most organic one, however, one must bear in
mind that there is no silver bullet for regulating something
as complex and multi-layered as the AI, and problems
may occur as well. Calibrating possible breaches of the
standards and converting them into a simplif‌ied four-tier
system can turn out to be over-simplistic and seize to ex-
ist after the f‌irst major scandal. This approach also raises
ethical issues, as ‘minimal’ or ‘acceptable’ risk suggests a
trade-off between the complicit needs.
The clarity of the regulation must appeal to the entrepre-
neurs, as opposed to the countries without sector-specif‌ic
restriction. Although the draft AI regulation is already
subject to public discussions until July 2021, the oppo-
sition among some businesses has already been f‌ierce
— developers and experts have repeatedly criticized the
‘loose’ def‌initions and huge costs that vendors would be
required to invest in order to comply with the new frame-
work (Ranchordas, 2021). Another controversy has been
12. EUR-Lex-52021PC0206. Proposal for a Regulation of the European Parliament and of the Council Laying Down Harmonised Rules on
Artificial Intelligence (Artificial Intelligence Act) and Amending Certain Union Legislative Acts (2021), Brussels, 21 April.
13. EUR-Lex-32016R0679. Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of
natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/
EC (General Data Protection Regulation). Official Journal L119, 4 May 2016, pp. 1–88.
exposed by human rights defenders: on the one hand, the
proposed framework intends to ban the ‘remote biometric
identif‌ication systems in publicly accessible spaces for the
purpose of law enforcement’. On the other hand, contrary
to the previously leaked document, the published version
adds an exception that opens the door to AI-powered
surveillance if allocating a suspect ‘of any criminal act
that carries a minimum three-year sentence’. Although
such action would require prior approval from court, it is
clear that this clause may provide a mass-scale violation
of human rights.
13
The experts are also concerned with the rather limited
scope of the AI regulation — namely, leaving behind the
lethal autonomous weapon systems (LAWS), civil liability,
intellectual property and competition issues arising from
the use of AI. Omitting the use of AI in military purposes
could be explained by the popular opinion on the AI tech-
nologies as rather an improvement for other operations,
than a way to develop completely new ones, which is
sometimes criticized as being ‘linear’ (Russell & Bohannon,
2015) if not misleading. Today the need for international
control over military AI became evident — such framework
must compliment ethics and legality with strategic goals
and consider the arsenal of AI-powered arms being largely
more extensive than the kinetic autonomous weapons sys-
tems (Schafer, 2016). The next point of critique, the lack of
new elaborations on the civil liability, may be explained by
the intention of the EU legislature to amend the Product
Liability Directive with respective provisions, already re-
f‌lected in the White Paper on AI (Schuett, 2019). However,
one should bear in mind that this instrument has been
developed in 1985 and does not fully match the technolog-
ical challenges of the modern era (Smuha, 2019). Finally,
the challenges of AI technologies to intellectual property
rights system are two-fold: f‌irstly, businesses seek to
patent AI solutions, which is not possible yet under the
European Patent Convention, unless it is as a ‘subgroup
of computer-implemented inventions’ (Stahl, 2021). Sec-
ondly, AI has already been denied twice for registration as
an inventor by the European Patent Off‌ice, based on the
https://idp.uoc.edu
Universitat Oberta de Catalunya
The European approach to building AI policy and governance: a haven for bureaucrats or innovators?
IDP No. 34 (December, 2021) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
8
2021, Olga Shumilo and Tanel Kerikmäe
of this edition: 2021, Universitat Oberta de Catalunya
lack of legal personality (Taube, 1961), which could also be
ref‌lected in the holistic proposal.
Anyhow, when contemplating the possible benef‌its and
shortcomings of the proposed AI regulation, we should
remember that delivering the proposal is only the f‌irst
step on the long road of legislative procedures, that takes
up to six years to make a comprehensive directive such
as GDPR to become fully enforceable all over the Union.
The European Parliament will hold several rounds of de-
bate, most importantly — trilogues, def‌ined as ‘informal
meetings attended by representatives of the Parliament
(rapporteur and, where appropriate, shadow rapporteurs),
the Council (chair of the working party and/or Commit-
tee of Permanent Representatives), and the Commission
(department responsible for the dossier and the Com-
mission’s Secretariat-General)’. After the Parliament and
the Council of the Union reach an agreement, the act
is adopted by both and published in the OJEU. Member
states are usually given another two years to adopt their
legislation and adjust the institutional framework, with
the total timeline of up to six years. This circumstance
puts a burden on the regulation creators, to keep an eye
on technological advances and constantly update the
regulation until it’s adopted in Brussels. Finally, Member
states may also put signif‌icant barriers for the regulation
to become a new reality. As proven by the GDPR case,
14
de-
spite the law already being directly enforceable across the
EU, some member states may still be reluctant to comply
with the new standards. It is also highly likely that the new
policy will trigger entrepreneurs and activists to discuss
its acceptability through the prism of local constitutional
and competition law (Troitino, 2014).
If enacted, the AI regulation can also inf‌luence interna-
tional technical standards on the use of AI. In general,
14. European Commission (2019). “Communication from the Commission to the European Parliament, the Council, the European Economic
and Social Committee and the Committee of the Regions-Building Trust in Human Centric Artificial Intelligence (COM-2019-168)”. In:
European Commission DS [online]. Available at: https://digital-strategy.ec.europa.eu/en/library/communication-building-trust-human-cen-
tric-artificial-intelligence
15. EUR-Lex-52020DC0064. Report From The Commission To The European Parliament. The Council And The European Economic And Social
Committee Report on the safety and liability implications of Artificial Intelligence, the Internet of Things and robotics [online]. Available
at: https://eur-lex.europa.eu/legal-content/en/TXT/?qid=1593079180383&uri=CELEX%3A52020DC0064
16. European Commission (2021). “Europe fit for the Digital Age: Commission proposes new rules and actions for excellence and trust in
Artificial Intelligence”. European Commission [online]. Available at: https://ec.europa.eu/commission/presscorner/detail/en/ip_21_1682
17. National Science and Technology Council (2016). “Preparing for the Future of Artificial Intelligence”. In: Obama White House [online].
Available at: https://obamawhitehouse.archives.gov/sites/default/files/whitehouse_files/microsites/ostp/NSTC/preparing_for_the_futu-
re_of_ai.pdf
the role of international standards in the EU can be de-
scribed as a soft law.
15
Although the standards remain
voluntary by nature, they provide building conf‌idence in
the market and make the compliance with a particular
directive auditable by an independent third party. As of
today, ISO/IEC JTC 1 SC 42, the joint subcommittee of the
ISO (International Organisation for Standardization) and
the IEC (International Electrotechnical Commission) has
already published eight standards on AI ecosystems, and
22 more are currently under development.
16
Hence, after
adopting the regulation, the European Commission may
either request one of the European Standards Organi-
sation: CEN (European Committee for Standardization),
CENELEC (European Committee for Electrotechnical
Standardization), or ETSI (European Telecommunications
Standards Institute) to develop a harmonised standard,
that would provide the industry stakeholders with a clear
compliance algorithm, or rely on the assumption that the
forthcoming international AI standards will incorporate
fundamental European values and a risk-based approach.
However, according to the latest report published by the
CEN-CENELEC Focus Group, the international standards
do not address all the societal and sovereignty concerns
that are relevant for the EU context, although further
technical cooperation with the ISO and IEC continues.
17
In addition to the presented regulation, the EU will take
further steps to regulate AI. Reportedly, after the core AI
regulation has gone through the debating phase, the Euro-
pean Commission is going to develop amendments to the
Product Liability Directive and the General Product Safety
Directive. Although it’s impossible to make the AI-related
legislation ‘future-proof’ given the nature of the technol-
ogy, the Commission will still try to avoid loopholes for
corporations, however these limitations should not scare
the big data giants off to other jurisdictions. Overall, it is
https://idp.uoc.edu
Universitat Oberta de Catalunya
The European approach to building AI policy and governance: a haven for bureaucrats or innovators?
IDP No. 34 (December, 2021) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
9
2021, Olga Shumilo and Tanel Kerikmäe
of this edition: 2021, Universitat Oberta de Catalunya
easy to predict the potentially catalyst impact of the AI
regulation on businesses, corporations erecting another
level of the compliance system and the inevitable confu-
sion among different stakeholders. While NGOs and unions
will be dissatisf‌ied with the act falling short of protecting
workers from AI-performed quality control or prof‌iling,
entrepreneurs will blame it as a compliance nightmare.
Conclusions
The advent of the AI era has been envisioned since the
1950s, but only f‌ive years ago the technology became om-
nipresent in our lives. The marketing power of the ‘AI’ brand
has maybe not yet reached its peak, though over 40% of
European self-proclaimed ‘AI start-ups’ reportedly have
nothing in common with the disruptive technology. Mean-
while, some corporations that already use it for recruiting
or attracting new users often wish to avoid responsibility
for possible racial discrimination or privacy abuse.
18. The phenomenon takes place when a jursidiction ’exports’ own strict standards to other countries due to a greater level of globalization,
as opposed to the ’race to the bottom’ trend (EUR-Lex-12007L/TXT).
19. National Science and Technology Council (2016). “The National Artificial Intelligence Research And Development Strategic Plan”. In:
Networking and Information Technology Research and Development Program [online]. Available at: https://www.nitrd.gov/pubs/natio-
nal_ai_rd_strategic_plan.pdf
As noted above, the prospective EU regulation will take at
least four years to be adopted, which will then show the
eff‌icacy of the chosen regulation model in real-life condi-
tions and maybe even replicate the ‘California effect’
18
of
the GDPR. In comparison to a more relaxed US legislation,
the consolidated EU bill may face long odds in the race for
fast economic revenue but protecting the Union against
mass surveillance and discrimination is a sustainable way
to stronger democracy and a higher level of public trust
in AI solutions, that could not be achieved by Member
States individually. It is also worth mentioning that the
EU has set an ambitious goal to build robust datasets and
computation infrastructure that would secure a unique
competitive advantage in capacity for research excel-
lence. As recently assessed by the European Parliament,
by 2030, a common AI regulatory effort can bring the EU
€294.9 billion in additional GDP and 4.6 million additional
jobs by 2030.
19
If adopted, the risk-based human-centric
framework will effectively inf‌luence EU allies and boost
sustainable economic growth across the globe.
https://idp.uoc.edu
Universitat Oberta de Catalunya
The European approach to building AI policy and governance: a haven for bureaucrats or innovators?
IDP No. 34 (December, 2021) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
10
2021, Olga Shumilo and Tanel Kerikmäe
of this edition: 2021, Universitat Oberta de Catalunya
References
AGAR, J. (2020). “What is science for? The Lighthill report on artif‌icial intelligence reinterpreted”.
In: British Journal for the History of Science, pp. 289-310 [online]. DOI:
https://doi.org/10.1017/
S0007087420000230
ARRIETA, A. B.; DIAZ RODRIGUEZ, N.; DEL SER, J.; BENNETOT, A.; TABIK, S.; BARBADO, A.; HERRE-
RA, F. (2020). “Explainable Artif‌icial Intelligence (XAI): Concepts, taxonomies, opportunities and
challenges toward responsible AI”. In: Information Fusion, pp. 82-115 [online]. DOI: https://doi.
org/10.1016/j.inffus.2019.12.012
BBC (2021). “EU artif‌icial intelligence rules will ban ‘unacceptable’ use”. In: BBC [online]. Available at:
https://www.bbc.com/news/technology-56830779
BLACK, J.; MURRAY, A. (2019). “Regulating AI and Machine Learning: Setting theRegulatory Agenda”.
In: European Journal of Law and Technology.
BÜTTNER, S. M.; LEOPOLD, L.; MAU, S.; POSVIC, M. (2015). “Professionalization in EU Policy-Making?
The topology of the transnational f‌ield of EU affairs”. In: European Societies, pp. 569-592 [online].
DOI: https://doi.org/10.1080/14616696.2015.1072229
CABRAL, T. S. (2020). “Liability and artif‌icial intelligence in the EU: Assessing the adequacy of the
current Product Liability Directive”. In: Maastricht Journal of European and Comparative Law, pp.
615–635 [online]. DOI: https://doi.org/10.1177/1023263X20948689
CEN-CENELEC (2020). “Road Map on Artif‌icial Intelligence (AI)”. In: CEN-CENELEC. Available at:
https://ftp.cencenelec.eu/EN/EuropeanStandardization/Sectors/AI/CEN-CLC_FGR_RoadMapAI.pdf
DE HERT, P.; PAPAKONSTANINOU, V.; KAMARA, I. (2016). “The cloud computing standard ISO/IEC 27018
through the lens of the EU legislation on data protection”. In: Computer Law and Security Review,
pp. 16-30 [online]. DOI: https://doi.org/10.1016/j.clsr.2015.12.005
DELPONTE, L. (2018). “European Artif‌icial Intelligence (AI) leadership, the path for an integrated
vision”. In: Eurocities [online]. Available at:
https://nws.eurocities.eu/MediaShell/media/Europe-
an_AI_study.pdf
DWIVEDI, Y. K. (2019). “Artif‌icial Intelligence (AI): Multidisciplinary perspectives on emerging chal-
lenges, opportunities, and agenda for research, practice and policy”. In: International Journal of
Information Management.
EUR-Lex-12007L/TXT. Treaty of Lisbon amending the Treaty on European Union and the Treaty estab-
lishing the European Community, signed at Lisbon (2007), Off‌icial Journal C306, 17 December, pp.
1–271
EUR-Lex-12012P/TXT. Charter of Fundamental Rights of the European Union. Off‌icial Journal C326, 26
October, pp. 391-407.
EUR-Lex-32001L0095. Directive 2001/95/EC of the European Parliament and of the Council of 3 De-
cember 2001 on General Product Safety, Off‌icial Journal L 11, 15 January 2002, pp. 4–17.EUR-Lex-
32006L0042. Directive 2006/42/EC of the European Parliament and of the Council of 17 May 2006
on machinery and amending Directive 95/16/EC (recast), Off‌icial Journal L 157, 9 June 2006, pp.
24–86.
EUR-Lex-32014L0053. Directive 2014/53/EU of the European Parliament and of the Council of 16 April
2014 on the harmonisation of the laws of the Member States relating to the making available on the
market of radio equipment and repealing Directive 1999/5/EC Off‌icial Journal L 153, 22 May 2014,
pp. 62–106.
https://idp.uoc.edu
Universitat Oberta de Catalunya
The European approach to building AI policy and governance: a haven for bureaucrats or innovators?
IDP No. 34 (December, 2021) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
11
2021, Olga Shumilo and Tanel Kerikmäe
of this edition: 2021, Universitat Oberta de Catalunya
EUR-Lex-32016R0679. Regulation (EU) 2016/679 of the European Parliament and of the Council of 27
April 2016 on the protection of natural persons with regard to the processing of personal data and
on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection
Regulation). Off‌icial Journal L119, 4 May 2016, pp. 1–88.
EUR-Lex-52018DC0237. European Commission. 2018, April 25. Communication from the Commission
to the European Parliament, the European Council, the Council, the European Economic and Social
Committee and the Committee of the Regions Artif‌icial Intelligence for Europe [online]. Available at:
https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=COM%3A2018%3A237%3AFIN
EUR-Lex-52020DC0064. Report From The Commission To The European Parliament. The Council And
The European Economic And Social Committee Report on the safety and liability implications of Ar-
tif‌icial Intelligence, the Internet of Things and robotics [online]. Available at: https://eur-lex.europa.
eu/legal-content/en/TXT/?qid=1593079180383&uri=CELEX%3A52020DC0064
EUR-Lex-52021PC0206. Proposal for a Regulation of the European Parliament and of the Council
Laying Down Harmonised Rules on Artif‌icial Intelligence (Artif‌icial Intelligence Act) and Amending
Certain Union Legislative Acts (2021), Brussels, 21 April.
European Commission (2019). “Communication from the Commission to the European Parliament, the
Council, the European Economic and Social Committee and the Committee of the Regions-Build-
ing Trust in Human Centric Artif‌icial Intelligence (COM-2019-168)”. In: European Commission DS
[online]. Available at:
https://digital-strategy.ec.europa.eu/en/library/communication-build-
ing-trust-human-centric-artif‌icial-intelligence
European Commission (2020). “White Paper on Artif‌icial Intelligence: a European approach to excel-
lence and trust”. In: European Commission [online]. Available at:
https://ec.europa.eu/info/publi-
cations/white-paper-artif‌icial-intelligence-european-approach-excellence-and-trust_en
European Commission (2021). “Annexes to the Communication from the Commission to the Europe-
an Parliament, the European Council, the Council, the European Economic and Social Committee
and the Committee of the Regions; Fostering a European approach to Artif‌icial Intelligence”. In:
European Commission DS [online]: Available at:
https://digital-strategy.ec.europa.eu/en/library/
coordinated-plan-artif‌icial-intelligence-2021-review
European Commission (2021). “Europe f‌it for the Digital Age: Commission proposes new rules and ac-
tions for excellence and trust in Artif‌icial Intelligence”. In: European Commission [online]. Available
at: https://ec.europa.eu/commission/presscorner/detail/en/ip_21_1682
European Union Agency for Cybersecurity (2020). “Artif‌icial Intelligence Cybersecurity Challenges”. In:
ENISA [online]. Available at:
https://www.enisa.europa.eu/publications/artif‌icial-intelligence-cy-
bersecurity-challenges
EVAS, T. (2020). “European framework on ethical aspects of artif‌icial intelligence, robotics and related
technologies”. In: European Parliament [online]. Available at:
https://www.europarl.europa.eu/
RegData/etudes/STUD/2020/654179/EPRS_STU(2020)654179_EN.pdf
INTRONA, L. D. (2016). “Algorithms, governance and governmentality: On governing academ-
ic writing”. In: Science, Technology & Human Values, pp. 17-49 [online]. DOI:
https://doi.
org/10.1177/0162243915587360
ISO/IEC JTC 1/SC 42 Artif‌icial intelligence [online] [2021]. Available at: https://www.iso.org/commit-
tee/6794475.html
JOHNSON, J. (2019). “The end of military-techno Pax Americana? Washington’s strategic responses to
Chinese AI-enabled military technology”. In: The Pacif‌ic Review, pp. 351-378 [online]. DOI:
https://
doi.org/10.1080/09512748.2019.1676299
https://idp.uoc.edu
Universitat Oberta de Catalunya
The European approach to building AI policy and governance: a haven for bureaucrats or innovators?
IDP No. 34 (December, 2021) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
12
2021, Olga Shumilo and Tanel Kerikmäe
of this edition: 2021, Universitat Oberta de Catalunya
KELLY, É. (2020). “EU struggles to go from talk to action on artif‌icial intelligence”. In: The Science
Business Network [online]. Available at: https://sciencebusiness.net/news/eu-struggles-go-talk-ac-
tion-artif‌icial-intelligence
LEE, N. (2018). “Detecting racial bias in algorithms and machine learning”. In: Journal of Informa-
tion, Communication and Ethics in Society, pp. 252-260 [online]. DOI: https://doi.org/10.1108/
JICES-06-2018-0056
MAAS, M. (2019). “How viable is international arms control for military artif‌icial intelligence? Three
lessons from nuclear weapons”. In: Contemporary Security Policy, pp. 285-311 [online]. DOI: https://
doi.org/10.1080/13523260.2019.1576464
MAEDCHE, A.; LEGNER, C.; BENLIAN, A.; BERGER, B.; GIMPEL, H.; HESS, T.; SÖLLNER, M. (2019). “AI-
based digital assistants”. In: Business & Information Systems Engineering, pp. 535-544 [online].
DOI: https://doi.org/10.1007/s12599-019-00600-8
MCCAULEY, L. (2007). “AI armageddon and the three laws of robotics”. In: Ethics and Information
Technology, pp. 153-164 [online]. DOI: https://doi.org/10.1007/s10676-007-9138-2
MEYER, D. (2018). “Most member states won’t be ready for GDPR”. In: IAPP [online]. Available at:
https://iapp.org/news/a/most-member-states-wont-be-ready-for-gdpr/
NATIONAL SCIENCE AND TECHNOLOGY COUNCIL (2016). “Preparing for the Future of Artif‌icial In-
telligence”. In: Obama White House [online]. Available at:
https://obamawhitehouse.archives.
gov/sites/default/f‌iles/whitehouse_f‌iles/microsites/ostp/NSTC/preparing_for_the_future_of_
ai.pdf
NATIONAL SCIENCE AND TECHNOLOGY COUNCIL (2016). “The National Artif‌icial Intelligence Re-
search And Development Strategic Plan”. In: Networking and Information Technology Research and
Development Program [online]. Available at:
https://www.nitrd.gov/pubs/national_ai_rd_strate-
gic_plan.pdf
NÍFHAOLÁIN, L.; HINES, A.; NALLUR, V. (2020). “Assessing the Appetite for Trustworthiness and the
Regulation of Artif‌icial Intelligence in Europe”. 28th Irish Conference on Artif‌icial Intelligence and
Cognitive Science-AICS. In: Research Repository UCD [online]. Dublin: UCD. Available at:
https://
researchrepository.ucd.ie/handle/10197/12396
PERRAULT, R.; SHOHAM, Y.; BRYNJOLFSSON, E.; CLARK, J.; ETCHEMENDY, J.; GROSZ, B.; NIEBLES, J. C.
(2019). “The AI Index 2019 Annual Report”. In: Human-Centered AI Institute [online]. Stanford: Stan-
ford University. Available at: https://hai.stanford.edu/sites/default/f‌iles/ai_index_2019_report.pdf
RANCHORDAS, S. (2021). “Experimental Regulations for AI: Sandboxes for Morals and Mores”. In: SSRN,
no. 7 [online]. Groningen: University of Groningen, Faculty of Law Research Paper. Available at:
SSRN:
https://ssrn.com/abstract=3839744.
DOI:
https://doi.org/10.5771/2747-5182-2021-1-86
RUSSELL, S.; BOHANNON, J. (2015). “Artif‌icial intelligence. Fears of an AI pioneer”. In: Science, vol.
349, no. 6245 [online]. DOI: https://doi.org/10.1126/science. 349.6245.252
SCHAFER, B. (2016). “Closing Pandora’s box? The EU proposal on the regulation of robots”. In: The
journal of the Justice and the Law Society, pp. 55-68. Brisbane: University of Queensland.
SCHUETT, J. (2019). “A Legal Def‌inition of AI”. In: SSRN [online]. DOI: https://doi.org/10.2139/
ssrn.3453632
SMUHA, N. A. (2019). “The EU Approach to Ethics Guidelines for Trustworthy Artif‌icial Intelligence”.
In: Computer Law Review International, pp. 97-106 [online]. DOI: https://doi.org/10.9785/cri-2019-
200402
https://idp.uoc.edu
Universitat Oberta de Catalunya
The European approach to building AI policy and governance: a haven for bureaucrats or innovators?
IDP No. 34 (December, 2021) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
13
2021, Olga Shumilo and Tanel Kerikmäe
of this edition: 2021, Universitat Oberta de Catalunya
STAHL, B. C. (2021). “EU is cracking down on AI, but leaves a loophole for mass surveillance”. In: The
Conversation [online]. Available at: https://theconversation.com/eu-is-cracking-down-on-ai-but-
leaves-a-loophole-for-mass-surveillance-159421
TAUBE, M. (1961). Computers and Common Sense. The Myth of Thinking Machines. New York: Columbia
University Press [online]. DOI: https://doi.org/10.7312/taub90714
THORNTON, R.; MIRON, M. (2020). “Towards the ‘Third Revolution in Military Affairs’ The Russian Mil-
itary’s Use of AI-Enabled Cyber Warfare”. In: RUSI Journal [online], pp. 12-21. DOI:
https://doi.org/1
0.1080/03071847.2020.1765514
TOTSCHNIG, W. (2020). “Fully autonomous AI”. In: Science and Engineering Ethics, no. 26, pp. 2473-
2485 [online]. DOI: https://doi.org/10.1007/s11948-020-00243-z
TROITINO, D. (2014). “The Single European Act-the Creation of the Interior Market in Europe”. In:
Journal on Legal and Economic Issues of Central Europe, pp. 8-14.
VINCENT, J. (2021). “EU outlines wide-ranging AI regulation, but leaves the door open for police sur-
veillance” [online]. In: The Verge. Available at:
https://www.theverge.com/2021/4/21/22393785/
eu-ai-regulation-proposal-social-credit-ban-biometric-surveillance-exceptions
VOGEL, D. (1995). Trading Up: Consumer and Environmental Regulation in a Global Economy. Cam-
bridge: Harvard University Press.
WACHTER, S.; MITTELSTADT, B.; RUSSELL, C. (2020). “Why Fairness Cannot Be Automated: Bridging
the Gap Between EU Non-Discrimination Law and AI”. In: SSRN Electronic Journal [online]. DOI:
https://doi.org/10.2139/ssrn.3547922
WANG, P. (2008). “What Do You Mean by “AI”?”. In: Frontiers in Artif‌icial Intelligence and Applications,
pp. 36 2-373.
WEARN, O. R.; FREEMAN, R.; JACOBY, D. M. (2019). “Responsible AI for conservation”. In: Nature Ma-
chine Intelligence, no. 1, pp. 72-73 [online]. DOI: https://doi.org/10.1038/s42256-019-0022-7
YU, R.; ALI, G. S. (2019). “What’s inside the Black Box? AI Challenges for Lawyers and Researchers”.
In: Legal Information Management, vol. 19, no.1, pp. 2-13 [online]. DOI: https://doi.org/10.1017/
S1472669619000021
ZAICHUK, M. (2018). “Patenting Artif‌icial Intelligence and Machine Learning Innovations in Europe”.
In: Jones Day [online]. Available at: https://www.jonesday.com/en/insights/2018/10/patenting-arti-
f‌icial-intelligence-and-machine-lear
ZAICHUK, M. (2020). “IP Protection of Artif‌icial Intelligence in Europe: Tailor-Made Solutions Re-
quired”. In: Jones Day [online]. Available at:
https://www.jonesday.com/en/insights/2020/04/
ip-protection-of-artif‌icial-intelligence-in-europe
Recommended citation
SHUMILO, Olga; KERIKMÄE, Tanel (2021). “The European approach to building AI policy and govern-
ance: a haven for bureaucrats”. IDP. Internet, Law and Politics E-Journal. No. 34. UOC [Accessed: dd/
mm/aa] http://dx.doi.org/10.7238/idp.v0i34.387744
https://idp.uoc.edu
Universitat Oberta de Catalunya
The European approach to building AI policy and governance: a haven for bureaucrats or innovators?
IDP No. 34 (December, 2021) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
14
2021, Olga Shumilo and Tanel Kerikmäe
of this edition: 2021, Universitat Oberta de Catalunya
The texts published in this journal, unless otherwise indicated, are subject to a Creative
Commons Attribution No Derivative Works 3.0 Spain licence. They may be copied, distribut-
ed and broadcast provided the the author, the journal and the institution that publishes
them (IDP. Revista de Internet, Derecho y Política; UOC) are cited. Derivative works are not
permitted. The full licence can be consulted on http://creativecommons.org/licenses/by-
nd/3.0/es/deed.es.
About the authors
Olga Shumilo
Tallinn University of Technology (TalTech)
olshum@taltech.ee
PhD student, writing her thesis on issues of legal framework. Doctoral student at Tallinn University of
Technology, researching the drivers and barriers behind the adoption of artif‌icial intelligence in judi-
ciary. She is also currently teaching Criminal and Constitutional Law, and has contributed to several
publications in the scope of EU Law and E-Governance. Olga has obtained her master’s degree in law
at the Yaroslav Mudryi National Law University (Kharkiv, Ukraine). In Ukraine, Olga has specialized in
Criminology, particularly animal welfare.
ORCID: https://orcid.org/0000-0001-7251-7947
Tanel Kerikmäe
Tallinn University of Technology (TalTech)
tanel.kerikmae@taltech.ee
Professor of European Law and director of Tallinn Law School at Tallinn University of Technology;
a board member of several high-ranked law journals and an author of more than 150 articles and
publications, Tanel is active as an expert for public and private institutions, international organiza-
tions and has been the keyexpert of EU in Central Asia. He has been contributing to the research of
European law and policies and created a research group of law & tech that received signif‌icant funding
from EU Commission (H2020, Jean Monnet Centre of Excellence), NATO, Folke Bernadotte Academy,
USAID, Estonian Research Council etc. He serves as a President of the Estonian branch of European
Community Studies Association andhas been teaching and supervising at several universities in and
outside Europe, including the Baltic States. Currently, prof. Kerikmäe is responsible for several inter-
national projects and related tointernational research cooperation networks. Moreover, his expertise
also includes the topics on EU legal policies, digital single market, automatisation, legal impediments
to new technologies, and augmented and artif‌icial intelligence in legal practice.
ORCID: https://orcid.org/0000-0002-5972-827X

VLEX utiliza cookies de inicio de sesión para aportarte una mejor experiencia de navegación. Si haces click en 'Aceptar' o continúas navegando por esta web consideramos que aceptas nuestra política de cookies. ACEPTAR