ShareDocView.com - Biometric & Behavioural Mass Surveillance in EU Member States

Report for the Greens/EFA in the
European Parliament
October 2021
BIOMETRIC
BEHAVIOURAL
MASS SURVEILLANCE
IN EU MEMBER STATES
&

TABLE
OF
CONTENTS
4 AUTHORS
5 ACRONYMS
7 EXECUTIVE SUMMARY
15 CHAPTER 1. Introduction
16 1.1 Objectives of the report
17 1.2 The international context
18 1.3 The European context
19 1.4 Four positions in the policy debates
20 1.5 Lack of transparency and the stifling of public debate
21 1.6 Scope and working definitions
22 1.7 Methodology
23 PART I: OVERVIEW OF EUROPEAN PRACTICES
25 CHAPTER 2. Technical overview
26 2.1 Remote Biometric Identification and classification: defining key terms
26 2.2 Detection vs recognition
26 2.3 Facial Recognition: verification/authentication vs identification
27 2.4 Forensic (ex-post) vs Live Facial Recognition
27 2.5 Other systems: gait recognition, emotion recognition
28 2.6 How does image-based remote biometric identification work?
31 2.7 Technical limits, problems, and challenges of facial recognition
36 CHAPTER 3. Overview of deployments in Europe
37 3.1 Authentication
41 3.2 Surveillance
44 3.3 Remote Biometric Identification
45 3.4 Conclusion
47 CHAPTER 4. Legal bases
48 4.1 EU Fundamental Rights Framework for the Right to Privacy and
the Right to Protection o Personal Data
50 4.2 EU Secondary Law: GDPR & LED
53 4.3 EU Soft law: Convention 108+
55 CHAPTER 5. Main political issues and debates
57 5.1 The emergence of remote biometric identification as a policy issue
57 5.2 Four positions in the policy debates
62 5.3 EU Commission Proposal on the Regulation for the Artificial
Intelligence Act

TABLE
OF
CONTENTS
65 PART II: CASE STUDIES
67 CHAPTER 6. Facial Recognition cameras at Brussels International Airport (Belgium)
68 6.1 The Zaventem pilot in the context of Face Recognition Technology in Belgium
69 6.2 Legal bases and challenges
71 6.3 Mobilisations and contestations
72 6.4 Effects of the technologies
74 CHAPTER 7. The Burglary Free Neighbourhood in Rotterdam (Netherlands)
75 7.1 Detection and decision-making in the “Burglary free neighbourhood” Fieldlab
83 CHAPTER 8. The Safe City Projects in Nice (France)
84 8.1 The various facets of the “Safe city” project in Nice
90 CHAPTER 9. Facial Recognition in Hamburg, Mannheim & Berlin (Germany)
91 9.1 RBI Deployments in Germany
96 9.4 Effects of the technologies: normalising surveillance
98 CHAPTER 10. The Dragonfly project (Hungary)
99 10.1 Remote Biometric Identification in Hungary
107 CHAPTER 11. Recommendations
110 REFERENCES
118 ANNEX: CASES
118 11.1 CJEU Decisions
118 11.2 ECtHR decisions
118 11.3 Decisions of National Courts

AUTHORS
Dr. Francesco Ragazzi (scientific coordinator) is an associate professor in Internation-
al Relations at Leiden University (Netherlands), an associated scholar at the Centre
d’Etude sur les Conflits, Liberté et Sécurité (France). He holds a PhD in Political Sci-
ence from Sciences Po Paris (France) and Northwestern University (USA). His research
interests include radicalisation, terrorism, and mass surveillance. His current research
project, Security Vision, funded by a European Research Council Consolidator Grant
analyses the politics of computer vision in the field of security. His work has been
published in numerous peer-reviewed journals and edited volumes. He serves on the
editorial board of the journals International Political Sociology, Citizenship Studies
and Cultures & Conflits. He has been consulted as an expert on issues of security by
the European Parliament, for whom he has co-authored several reports, the Council
of Europe and the French Senate.
Dr. Elif Mendos Kuskonmaz is a lecturer at the School of Law at the University of
Portsmouth. She holds a Master’s Degree in Public Law from Istanbul University, and
an LLM in Public International Law and a PhD from Queen Mary University of London.
She researches on surveillance measures and the nexus with the right to privacy
and data protection. Elif is also a registered lawyer with the Istanbul Bar Association.
Ildikó Z Plájás is a post-doctoral researcher at the Institute of Political Science, Leiden
University. She has studied anthropology and cultural studies in Romania and Hun-
gary, later graduating in Visual Ethnography at Leiden University, the Netherlands.
She is currently completing her PhD at the University of Amsterdam. Her research
examines how visual technologies in governance enact certain groups of people as
“racial others” in Europe.
Ruben van de Ven is a PhD candidate in Political Science at the Institute of Political
Science, Leiden University. His PhD project studies the ethical and political implica-
tions of surveillance algorithms that order human gestures. Since graduating from
the Master in Media Design programme at the Piet Zwart Institute, he has researched
algorithmic politics through media art, computer programming and scholarly work. He
has focused on how the human individual becomes both the subject of and input into
machine learning processes. Earlier artistic work on the quantification of emotions
examined the transformation of humanistic concepts as they are digitised. His work
has been presented at both art exhibitions and academic conferences.
Dr Ben Wagner is an assistant professor at the Faculty of Technology, Policy and
Management at TU Delft, where his research focuses on technology policy, human
rights and accountable information systems. He is Associate Faculty at the Complexity
Science Hub Vienna and a visiting researcher at the Human Centred Computing Group,
University of Oxford. He previously worked at WU Vienna, TU-Berlin, the University of
Pennsylvania and European University Viadrina. He holds a PhD in Political and Social
Sciences from the European University Institute in Florence

ACRONYMS
ABIS Automated Biometric Identification
Systems
ACLU American Civil Liberties Union
ADM Automated Decision-Making
(System)
AFIS Automated Fingerprint
Identification System
AI Artificial Intelligence
ANPR Automated Number Plate
Recognition
API Application Programming
Interface
AWS Amazon Web Services
BDAS Biometric Data Processing
System
BDSG Federal Data Protection Act
(Germany)
BKA Federal Criminal Police Office
(Germany)
BKK Centre for Budapest Transport
(Hungary)
BPI Public Investment Bank
(France)
BPOL German Federal Police
CATCH Central Automatic TeChnology for
Recognition of Persons
(Netherlands)
CBIS Central Biometric Information
System (Czechia)
CCTV Closed Circuit Television
CGT General Labour Confederation
(France)
CJEU Court of Justice of the European
Union (EU)
CNIL National Commission for
Informatics and Freedoms
(France)
COC Supervisory Body for Police
Information (Belgium)
CoE Council of Europe
COCO Common Objects in Context
(Dataset)
COVID Coronavirus Disease
CSU Centre for Urban Supervision
(France)
DEP Digital European Program

DITSS Dutch Institute for Technology,
Safety & Security
DPA Data Protection Authority
EC European Commission (EU)
ECtHR European Court of Human Rights
EDE Criminal identification database
(Austria)
EDPB European Data Protection Board (EU)
EDPS European Data Protection
Supervisor (EU)
EDS European Data Strategy
EEA European Economic Area
EPP European People’s Party
EU European Union
FRA Fundamental Rights Agency
(EU)
FRT Facial Recognition Technology
FRVT Face Recognition Vendor Test
GDPR General Data Protection Regulation
(EU)
HCLU Hungarian Civil Liberties Union
(Hungary). See “
HD High Definition
HDR Habitoscopic Data Register
HKR Home Quarantine App
(Hungary)
IARPA Intelligence Advanced Research
Projects Agency (USA)
ID Identification
IFRS Interpol Facial R
IKSZR Integrated Traffic Management and
Control System (Hungary)
INCLO International Network of Civil
Liberties Organisations
INPOL Criminal Case Management System
(Germany)
KAK Governmental Data Centre
(Hungary)

ACRONYMS
KDNP Christian Democratic People's
Party
(Hungary)
LED Law Enforcement Directive
(EU)
LFP Law on the Function of Police
(Belgium)
LGBTQ Lesbian, Gay, Bisexual,Transgender,
Queer
LIDAR Light Detection and Ranging
LPA Airport Police (Belgium)
LQDN La Quadrature du Net (France)
GMO Genetically Modified Organism
MIT Massachusetts Institute of
Technology
MRAP Movement against racism and for
friendship between peoples
(France)
NAIH Hungarian National Authority for
Data Protection and Freedom of
Information
NBIS National Biometric Identification
System (Romania)
NGO Non-Governmental
Organisation
NIST National Institute of Standards and
Technology (USA)
NISZ National Infocommunication
Services (Hungary)
PARAFE Rapid passage at the external
borders (France)
PPM Pixels Per Meter
RBI Remote Biometric
Identification
RETU Registered persons identifying
features database and Aliens
database (Finland)
RGB Red, Green, Blue
SIS Schengen Information Sys-
tem
SSNS Secret Service for National Securi-
ty
(Hungary)
TAJ Criminal case history database
(France)
TASZ Hungarian Civil Liberties
Union
TELEFI Towards the European Level
Exchange of Facial Images
(EU Project)
UAVG GDPR Implementation Act
(Germany)
UK United Kingdom
UN United Nations
UNHRC United Nations Human Rights
Council
US(A) United States of America
VGG Visual Geometry Group
(Dataset)
VMD Video motion detection
VOC Visual Object Classes
(Pascal VOC)
YOLO You Only Look Once
(Algorithm)

7
EXECUTIVE SUMMARY
CHAPTER 1: INTRODUCTION
The aim of this report is to establish a proble-
matised overview of what we know about what
is currently being done in Europe when it co-
mes to remote biometric identification (RBI),
and to assess in which cases we could poten-
tially fall into forms of biometric mass surveil-
lance.
Private and public actors are increasingly de-
ploying “smart surveillance” solutions inclu-
ding RBI technologies which, if left uncheck-
ed, could become biometric mass surveillance.
Facial recognition technology has been the
most discussed of the RBI technologies.
However, there seems to be little understan-
ding of the ways in which this technology
might be applied and the potential impact of
such a broad range of applications on the fun-
damental rights of European citizens.
The development of RBI systems by authori-
tarian regimes which may subsequently be
exported to and used within Europe is of con-
cern. Not only as it pertains to the deploy-
ments of such technologies but also the lack
of adequate insight into the privacy practices
of the companies supplying the systems.
Four main positions have emerged among po-
litical actors with regard to the deployments
of RBI technologies and their potential impact
on fundamental rights: 1) active promotion 2)
support with safeguards; 3) moratorium and
4) outright ban.
CHAPTER 2: TECHNICAL OVERVIEW
The current market of RBI systems is over-
whelmingly dominated by image-based pro-
ducts, at the centre of which is facial recogni-
tion technology (FRT). Other products such as
face detection and person detection techno-
logies are also in use.
FRT is typically being deployed to perform two
types of searches: cooperative searches for
verification and/ or authentication purposes,
and non-cooperative searches to identify a
data subject. The former involves voluntary
consent from the data subject to capture their
image, while the latter may not.
Live facial recognition is currently the most
controversial deployment of FRT: Live video
feeds are used to generate snapshots of indi-
viduals and then match them against a data-
base of known individuals – the “watchlist”.
Other RBI technologies are being deployed
though their use at present is marginal com-
pared to FRT, these include gait (movement),
audio, and emotion recognition technologies,
amongst others.
A better understanding of the technical com-
ponents and possible usage applications of

8
image-based RBI technologies is needed in
order to assess their potential political impli-
cations.
RBI technologies are subject to technical chal-
lenges and limitations which should be consi-
dered in any broader analysis of their ethical,
legal, and political implications.

CHAPTER 3: OVERVIEW OF DEPLOY-
MENTS IN EUROPE
Current deployments of RBI technologies
within Europe are primarily experimental and
localised. However, the technology coexists
with a broad range of algorithmic processing
of security images being carried out on a scale
which ranges from the individual level to what
could be classed as biometric mass surveillan-
ce. Distinguishing the various characteristics
of these deployments is not only important
to inform the public debate, but also helps to
focus the discussion on the most problematic
uses of the technologies.
Image and sound-based security applications
being used for authentication purposes do not
currently pose a risk for biometric mass sur-
veillance. However, it should be noted that an
alteration to the legal framework could increa-
se the risk of them being deployed for biome-
tric mass surveillance especially as many of
the databases being used contain millions of
data subjects.
In addition to authentication, image and
sound-based security applications are being
deployed for surveillance. Surveillance app-
lications include the deployment of RBI in
public spaces.
Progress on two fronts makes the developme-
nt of biometric mass surveillance more than
a remote possibility. Firstly, the current cre-
ation and/or upgrading of biometric databa-
ses being used in civil and criminal registries.
Secondly, the repeated piloting of live-feed
systems connected to remote facial and bio-
metric information search and recognition al-
gorithms.
CHAPTER 4: LEGAL BASES
The use of biometric tools for law enforcement
purposes in public spaces raises a key issue of
the legal permissibility in relation to the col-
lection, retention and processing of data when
considering the individual’s fundamental
rights to privacy and personal data protection.
When viewed through this lens, RBI technolo-
gies could have a grave impact on the exercise
of a range of fundamental rights.
The deployment of biometric surveillance in
public spaces must be subject to strict scru-
tiny in order to avoid circumstances which
could lead to mass surveillance. This includes
targeted surveillance which has the potenti-
al for indiscriminate collection of data on any
persons present in the surveilled location, not
only that of the target data subject.
The normative legal framework for conduc-
ting biometric surveillance in public spaces
can be found in the EU secondary legislation
on data protection (GDPR and LED). The use of
biometric data under this framework must be
reviewed in light of the protection offered by
fundamental rights.
The European Commission’s April 2021 propo-
sal on the Regulation for the Artificial Intelli-
gence Act aims to harmonise regulatory rules
for Member States on AI-based systems. The
Proposed Regulation lays out rules focused on
three categories of risks (unacceptable, high,
and low/ minimal risk) and anticipates cove-
ring the use of RBI systems. It also aims to
compliment the rules and obligations set out
in the GDPR and LED.

CHAPTER 5: POLITICAL DEVELOPME-
NTS AND MAIN ISSUES OF CONTENTION
Four main positions on RBI systems have
emerged among political actors as a result of

9
both technical developments in the field and
early legislative activity of EU institutions: 1)
active promotion 2) support with safeguards;
3) moratorium and 4) outright ban.
Those who are in favour of support with safe-
guards argue that the deployment RBI tech-
nologies should be strictly monitored because
of the potential risks they pose, including the
potential danger of FRT, for example, to contri-
bute to the further criminalisation or stigma-
tisation of groups of people who already face
discrimination.
The European Parliament passed a resolution
on artificial intelligence in January 2020 in
which they invite the Commission “to assess
the consequences of a moratorium on the use
of facial recognition systems”. If deemed ne-
cessary, such a moratorium could impact some
existing uses of FRT including its deployment
in public spaces by public authorities.
A number of EU and national NGOs have called
for an outright ban on the use of RBI with some
arguing that the mass processing of biometric
data from public spaces creates a serious risk
of mass surveillance that infringes on funda-
mental rights.
The European Commission’s legislative propo-
sal for an Artificial Intelligence Act (EC 2021)
is both a proposal for a regulatory framework
on AI and a revised coordinated plan to sup-
port innovation. One feature of the act is the
establishment of risk-dependent restrictions
which would apply to the various uses of AI
systems.
CASE STUDIES
CHAPTER 6: FACIAL RECOGNITION CA-
MERAS AT BRUSSELS INTERNATIONAL
AIRPORT (BELGIUM)
Belgium is one of two European countries that
has notyet authorised the use of FRT, however,
law enforcement is strongly advocating for its
use and the current legal obstacles to its im-
plementation are unlikely to hold for very long.
In 2017, unbeknownst to the Belgian Supervi-
sory Body for Police Information (COC), Brus-
sels International Airport acquired 4 cameras
connected to a facial recognition software
for use by the airport police. Though the COC
subsequently ruled that this use fell outside
of the conditions for a lawful deployment, the
legality of the airport experiment fell into a le-
gal grey area because of the ways in which the
technology was deployed.
One justification for the legality of the airport
experiment from the General Commissioner of
Federal Police was to compare the technologi-
cal deployment to that of the legal use of other
intelligent technologies such as Automated
Number Plate Recognition (ANPR). Although
this argument was rejected at the time, such
a system could be re-instated if the grounds
for interruption are no long present in the law.
There is an emerging civil society movement in
Belgium contesting the legitimacy of remote
biometric identification. However, the amend-
ments to the Police Act permitting the use of
real-time smart cameras by the police in car-
rying out their administrative and judicial du-
ties, and recent declarations of the Minister of
Interior seems to point in the direction of more
acceptance for remote biometric surveillance.
CHAPTER 7:THE BURGLARYFREE NEIG-
HBOURHOOD IN ROTTERDAM (NETHER-
LANDS)
The Fieldlab Burglary Free Neighbourhood is a
public-private collaboration with two aims: to
detect suspicious behaviour and to influence
the behaviour of the suspect. While the system
of smart streetlamps does collect some image
and sound-based data, it does not record any
characteristics specific to the individual.
From a legal perspective, there is a question
as to whether or not the data processed by
the Burglary Free Neighbourhood programme

Drop a Comment for
Full Document
Free
Signup at
ShareDocView.com

ShareDocView.com - Biometric & Behavioural Mass Surveillance in EU Member States

More Related Content

Similar to ShareDocView.com - Biometric & Behavioural Mass Surveillance in EU Member States

More from ShareDocView.com

Recently uploaded

ShareDocView.com - Biometric & Behavioural Mass Surveillance in EU Member States