Welcome address of the EU Commissioner of Home Affairs (video)
Welcome Address by CEPOL Executive Director
Complementary Coffee and Refreshment
The world around policing is changing as radically as it was in 1828 when Sir Robert Peel – founder of London’s Metropolitan Police Service – argued that Britain had outgrown her policing institutions and required a fundamentally “new mode of protection”. Today, the digital revolution, together with environmental crisis and complex social change, is transforming society, crime, and public safety, and posing challenges for governments and police agencies that are at least as profound.
In March 2022 the Police Foundation, the UK’s leading, independent policing think tank, published the final report of its Strategic Review of Policing in England and Wales, a two-year programme of work to consider the policing arrangements needed to meet the public safety challenges of the mid-21st century. In it, we argue that the challenge of capacity can only be met by creating an explicit public safety system, primarily focussed on prevention, that stretches well beyond traditional law enforcement, but within which police must play unique and vital role. We conclude that, to fulfil that role, policing needs to develop and maintain a set of core strategic capabilities: the ability to work co-operatively with the public, to make best use of technology, develop new skills and promote learning within its workforce, ensure its people are resilient and stay well, and demonstrate strong leadership. Finally, we consider the organisational platform best able to deliver these capabilities effectively and efficiently.
The keynote will:
• Introduce the Police Foundation and the Strategic Review, describing its objectives, remit, and methods.
• Describe how technological change and other forces are reshaping the public safety context in ways that demand a fundamental redesign of our policing arrangements.
• Make the case for a realignment towards a preventive public safety system, especially in relation to technologically enabled crimes, and consider the police role within it.
• Consider how police can secure legitimacy and public consent in the digital age.
• Focus on the technological capabilities and skills profile required in coming decades, and the shift to a learning culture needed to ensure these are continually developed.
The Dutch National Police had been struggling to improve crime fighting. A lot of challenges (personnel, knowledge, ICT and a lot of fragmentation were at the root of the problem. In response, a new concept of crime fighting, where speed and intelligence are key, has been developed. This concept is made operational via eight experiments and living labs. These developments are properly evaluated in terms of their contribution to crime fighting. Speeding up forensic analysis and giving direction to the inquiry can only be done by standardizing, organizing and making the process digital. Imagine that we speed up the analysis of fingerprints, facial recognition, DNA and drugs. This is done in a collaboration between the Police, the District Attorney’s Office, and the commercial market. The presentation will elaborate on the Dutch lessons learned over the last few years, ‘that speed matters’ at the centre of it. The next steps will be to scale up the new ways of working.
Complementary Light Lunch
Technological innovations such as digitalisation have an increasing role in our society. This development is also reflected in police work. In particular, the access to information at a global scale has increased the international character, adaptivity and fluidity of criminal organisations. As such, there is a pressing need to better understand the evolving nature of these organisations and their associated modus operandi. While digitalisation enables access to lots of information and yields information overload challenges, developments in Artificial Intelligence (AI) offer new opportunities to tackle these challenges. In particular, they provide support in the automatic extraction and analysis of unstructured sources of information to efficiently make sense of large amounts of textual information sources. In this paper we will explore the potential and challenges of AI methods to extract criminal modus operandi from unstructured open text sources, like law court sentences. Such open text sources are a reliable information source that includes detailed validated information on the criminal activities and the modus operandi evolution in a given country. The application of this approach offers an alternative to the examination of classified police information and it also facilitates cross-country comparisons.
The inherent complexity of modus operandi and the unstructured character of law court sentences yield the need to align and structure the modus operandi question with particular text mining methods. Specifically, we propose a step-wise approach to analyse automatic extraction of modus operandi related problems via exploration, detection and categorisation analysis. This decomposition enables to align these problems to specific functions of text-mining or machine learning methods, such as similarity detection, clustering or entity recognition. Using practical examples we demonstrate how this approach enables to automatically extract relevant information from court cases for analysing modus operandi evolution in time.
Co-authors: Koen van der Zwet, Joris Westerveld
This paper aims at studying the shortcomings and strengths of the French Law Enforcement administration in the use of E-learning with a focal point on the Police Nationale.
The evolution of law enforcement knowledge, techniques, and materials, make continuing education for the profession more essential than ever. The rise of digital technology in our societies has simultaneously changed the forms of education and the means of accessing this knowledge.
Yet, for a long time, the French Police Nationale refused to take the turn of digital progress in the training it offered to cadets but also to experienced officers on the grounds that it was inefficient and approximative. Reputedly monolithic and hard to modernize, the LE system in France was reluctant to the idea of using a tool they saw as unprofessional and unserious.
The Pandemic shattered their certainties. Unable to give in-person trainings but having to ensure the continuity of curriculums, the French Police Nationale training department had to think out of the box. A cumbersome process ensued with the construction of a new online training structure, the hiring of new digital experts and education-specialised civilians and the funding for new digital tools to implement quality courses. New training ideas emerged with these new recruits and new tools. Soon, major e-learning projects were achieved among which the cadet training curriculum and language program are now the best examples.
These, however, did not come without challenges for trainers and trainees alike. Trainers were not used to online tools. Where some welcomed novelty, others rejected it entirely. Those who used those tools faced numerous hindrances related to technology, internet connection, organisation, activity creation but also student motivation. Trainees were facing the same technological and organisational issues. Sometimes unmotivated by the lack of interactivity of courses or monotony of self-paced activities, some were even skipping online classes. These issues were eventually considered and a whole new program was set to give Police Nationale trainers an understanding of Online e-learning along with a willingness to bring more interactivity in online classes with the use of more powerful tools.
Digitalisation, one of the key elements addressed by CEPOL in law enforcement training, is carried out based on the continuing and emerging technological innovations that needs to be given the highest priority across the European law enforcement community.
The new European Union Strategic Needs Assessment (EU-STNA), defines the strategic EU-level training priorities of law enforcement officials for the next 4-year cycle, 2022-2025, in line with the EMPACT priorities, emphasizing the importance of digital skills and use of new technologies, as one of the main horizontal aspects that should be addressed in all training activities.
Cyber-attacks, had the highest prioritising, as EU training need, within the Member States, indicating more than 7600 officials that would need to be trained. Law enforcement and judiciary authorities would need further awareness raising regarding cyber security, cyber-enabled and cyber-dependent crime, but also further improvement in dealing with e-evidence and international cooperation mechanisms.
Taking into consideration the deliverables of the EU-STNA process, CEPOL has further launched a structured training needs analysis in 2021, the OTNA on Digital skills and the use of new technologies, in order to define the training portfolio addressing digitalisation of law enforcement for 2023-2025. Amongst most relevant training topics of the responding countries, we can highlight the Digital investigations, use of new technologies and digital forensics, that would need to be included in law enforcement training activities.
Keywords: digitalisation, digital skills, new technologies, law enforcement training, EU priorities
The use of artificial intelligence / machine learning models in law enforcement is becoming a necessity. A multitude of use cases range from video analytics of public surveillance systems with their facial recognition, object detection, and traffic safety control, to more specific pattern analysis of data used to uncover criminal enterprises. The common element is the analysis of the expanding amounts of available data to uncover key features, i.e. finding “needles in the haystack”.
The research project KISTRA has developed an application in collaboration with German LEAs to process a set of media data points consisting of images, videos and text using AI models for facial recognition, object detection and text analysis. The results are displayed in a unified user interface with a filterable map, highlighting the features of interest found. A unified user interface shows the results of the algorithms clearly and transparently. It focuses on functionalities that enable the user to further filter the dataset based on the objects, faces, and keywords in a configurable way. This enables the analyst to narrow down the dataset to select few data points of interest, and if needed to sort these data into different collections for export, further evaluation and integration with external tools.
We highlight the process and pitfalls of creating and deploying machine learning models and the need to curate non-biased, well-balanced training data, especially in light of the EU Artificial Intelligence Act proposal. The performance of the application is shown in a practical scenario, evaluating up to 100.000 media items with facial recognition, object detection and OCR and text analysis. Our use case shows that the time the analyst spends on finding relevant information can be reduced by more than 90% on average.
Co-author: Jens Elsner
The Bavarian Police training aims to equip all 750 police teachers and 4.000 police officer trainees with officially approved police tablet PCs and smartphones by the end of 2025. Following a phased approach, teachers and trainees are being issued with tablet PCs and smartphones. Furthermore, all classrooms will be equipped with interactive whiteboards by the end of 2021.
Overall, there is no doubt that the use of digital devices is undoubtedly important and necessary in order a) to prepare the police officer trainees for their future work as police officers and b) to support their learning process. However, the question arises as to how exactly they benefit from using those digital devices.
To get a first insight on the impact of digital devices on learning interest and engagement of police officer trainees in the classroom, as well as on their academic performance in general, the Bavarian Police training conducted a digital pilot project with a single unit of 100 police officer trainees and 20 police teachers and trainers for 21 months (December 2019 to August 2021).
The findings show that the different digital devices have different impacts on the learning behaviour as well as on the academic performance of the police officer trainees. Above all, tablet PCs and interactive whiteboards have shown to improve learning behaviour. Furthermore, the findings show that digital devices which are not used regularly do not improve the classroom behaviour of the police officer trainees or even worsens things slightly.
The study suggests several practical implications for the further implementation of digital devices such as the necessity of training the police personnel, developing new didactic teaching methods as well as new teaching formats and, last but not least, the necessity to realize and accept that the new generation of police officer trainees want to be taught through digital devices and digital teaching material.
UNCOVER – a joint international initiative funded by the EC under the H2020 Research & Innovation program – was launched in May 2021. UNCOVER is end-user-driven and responds to real needs of Law Enforcement Agencies to detect hidden information in innocent looking data (i.e., steganalysis). End-users are involved throughout the project cycle: from the analysis of user requirements and tools development through evaluation. Regular feedback cycles will ensure that the developed solutions can be integrated into the daily criminal investigation pipeline of LEAs. Besides its research activities, UNCOVER has a strong training dimension and over the course of the project, UNCOVER will establish a comprehensive train-the-trainers program for LEAs and forensic institutions, integrating all theoretical findings and solutions developed during the project.
UNCOVER sets out to outperform available steganalysis solutions in terms of performance, usability, operational needs, privacy protection, and chain-of-custody considerations. Project objectives are:
1. Advance the scientific state-of-the-art in steganalysis.
2. Develop, test and evaluate solutions for real-life LEA problems, based on their end-user requirements and expectations.
3. Analyze and implement proper solutions for relevant security, ethical, legal and privacy related concerns.
4. Develop and implement interdisciplinary technical capabilities, capacity and community building.
5. Engage and collaborate with any other relevant ongoing projects and initiatives.
Specific objectives of UNCOVER are to:
1) CONDUCT a detailed analysis about the various aspects of the needs and requirements of LEAs for detecting and investigating steganography.
2) CONSOLIDATE relevant information about existing steganographic tools and centralize this information in an intuitive database for LEAs.
3) IMPROVE existing methods for operational steganalysis in digital media workflows.
4) IMPLEMENT a platform for the integration of steganalysis detection tools.
5) DEMONSTRATE the steganographic detection capabilities with realistic test cases and scenarios delivered by LEAs.
6) ENSURE the obtained results are admissible in European court rules. Assist LEAs with adopting the technology in their daily work.
7) PROVIDE a comprehensive training program for LEAs and forensic institutes by providing in-house training.
8) DISSEMINATE outcomes, communicate the project and prepare an exploitation and sustainability plan.
CONNEXIONs is a H2020 EU funded project that aims to develop and demonstrate next-generation detection, prediction, prevention, and investigation services that are based on the multidimensional integration and correlation of heterogeneous multimodal data, as well as on the delivery of pertinent information to various stakeholders in an interactive manner tailored to their needs, including through augmented reality (AR) and virtual reality (VR) environments.
CONNEXIONs will present its achievements related to its use case on improving crime scene investigation and training through 3D reconstruction and VR applications.
For 3D reconstruction, CONNEXIONs leverages the emergence of high-precision and low-cost devices capable of scanning scenes or objects in 3D, proven a reliable means for the digital documentation of the scene. To this end, CONNEXIONs has developed a methodology and best practices for acquiring data for 3D reconstruction across a wide range of scales in indoor and outdoor scenes, emplacing the applicability of each technique in a wide range of situations, ranging in type and size. The application of each reconstruction method is considered in this context and compared with respect to additional constraints, such as time availability and simplicity of operation of the corresponding scanning modality.
Regarding the VR applications developed by the project, CONNEXIONs uses state-of-the-art VR technology for the purposes of training and post-event analysis. VR allows users to intuitively explore real-world environments, represented by realistic 3D reconstructions, and can embed information from multiple sources into a single scene, facilitating the intuitive understanding of the event or a crime scene. To facilitate understanding of the process of events, data, such as videos and locations. can be played back synchronously and in their spatial context. The analysis can be further enhanced by adding certain documents (e.g., photos of the event or crime scene). Moreover, the VR Training Application is a device-agnostic application that offers a variety of functionalities to create a wide range of different scenarios within a single reconstructed environment, complemented with a visual analytics tool providing additional insights into trainee’s actions and walking paths within the environment.
Cyprus Police, like all police organizations had to face the pandemic situation immediately and effectively. The challenges and difficulties were unprecedented both in the operational and the technical sector. The staff, trainers and trainees of the Cyprus Police Academy were called upon to adjust to the new circumstances in fulfilling their mission of educating, training and preparing the new recruits to become police officers.
Regardless of the considerable hesitations expressed in the past regarding distance learning, the digitalization of education at the Cyprus Police Academy became a fact, beginning with the online training of new recruits. If this change was not introduced, the training of new recruits would have to stop, at a crucial period where there was an urgent need for more police officers.
The purpose of this proposed research study is to focus on the commencement of the digitalization training of law enforcement officials in Cyprus, as a result of the pandemic measures, by identifying the main challenges, needs and problems faced, as well as the readiness and capability level of those involved in the on-line training and the use of digital tools.
To do so, a survey is to be carried out among trainers and staff of the Cyprus Police Academy. Specifically, qualitative data will be collected using a semi-structured questionnaire, which will be completed through personal interviews with at least 12 members.
This research study, which is the first of its kind at the Cyprus Police Academy, is expected to provide outcomes and suggestions in relation to the digitalization of police education not only in Cyprus but other law enforcement Organizations as well and form the basis for further research in applying good practices and modern training methods.
Nowadays police authorities, implement computational techniques like machine learning algorithms to forecast criminal details such as crime scenes, or personal characteristics of potential perpetrators. In Greece however, Artificial Intelligence and its potential are not widely explored. In this paper, we collected unstructured data from press releases of the Hellenic Police from 2015 to 2020, and then we manually constructed the dataset. More specifically, variables like the time of the investigation of each case, the place of the arrest, along with nationality, age, type of the offense, and confiscated items were included in the dataset. In the case of a conflict between collected values, the press release in question was omitted from the data. Furthermore, we matched the place of the arrest with its corresponding postal code and added some key statistics from the Hellenic Statistical Authority, namely the population of each area, the marital status of those living in every neighborhood, and their employment (i.e. merchants, farmers, employees, retirees).
For the experiments, we used tree-based machine learning models from the NLTK Library to predict between different kinds of crime. We conducted two experiments using Python, the first aimed at classifying burglary, theft, and fraud and the second was a binary classification between burglary and theft. The training set consisted of 70% of the data and the rest 30% was used for testing, also the F1-score was implemented for the evaluation of the model. The findings showed that the binary classification produced better results, with the model being able to predict between burglary and theft with 75% accuracy. This is the first attempt to our knowledge that implements predictive models to forecast crime in Greece, and our results suggest that this is a promising approach that might assist in daily policing. We consider that our experiment would be a start for the Hellenic Police to know how to predict and analyze crime, starting from the data interpretation and ending with the algorithmic models.
Complementary Coffee and Refreshment
Presentation of the open access tool; a 4-module e-learning course for International Police Advisors developed by the Norwegian Police University College, a delivery to the Horizon 2020 EU-funded research project "Community based Policing and Post-conflict Police Reform" (ICT4COP).
The course and the supplementary downloadable "Work Book" are aimed for international Police Advisors. But the course will also be beneficiary for military personnel, international civilian actors as well as civilian organizations and/or individuals from host society engaged in a police reform.
The authors, working a project mapping how law conceptualizes and operationalizes race, ethnicity and nationality, provide an assessment of the triadic relationship between law, law enforcement practices and science. The paper begins by providing an overview of the obstacles, challenges and controversies in the legal institutionalization and operationalization of ethnic/racial/national group affiliation. Subsequently, the paper turns to the assessment of how “objective” criteria, data and constructions provided by science and biotechnology translate into the legal discourse and more specifically law enforcement practice. The focus here will be on race-focused forensic datasets and how ever so digital law enforcement registries can operationalize ethno-racial data for profiling and big data analysis. The paper will use the case study and example of the legal framework and practice in Hungary.
The term ‘human factors’ covers extra-legal, social, psychological, institutional and organisational aspects affecting the behaviour of citizens, including the fight against financial crime by competent authorities. In addition to the complexities imposed by different legal approaches across EU Member States and the limited resources available for countering financial crime, managing and converting big data into meaningful law enforcement action have also become major challenges for LEAs. As the recent Pandora Paper demonstrate, it is not only the volume and type of data which can overwhelm the authorities but also the sheer number of people, companies, business operations and assets that are scattered across the globe. This is where AI informed data analytics can be of use in analysing the vast amount of structured and unstructured data as well as connecting the dots between entities involved in a given scenario. Based on real case studies, this paper presents the potential benefits as well as challenges in adopting AI solutions to financial crime investigation models. In doing so, it offers a critical overview of the implications of some of the salient aspects of natural language process, entity extraction, correlation analysis, visualization & crowd knowledge. The paper also presents a some of the core legal and ethical principles that shall inform the development AND use of AI in law enforcement practices in Europe.
Instructional strategies in many operative fields, including law enforcement, have reached a high level of complexity due to dynamically changing task environments and the introduction of different technologies to help the users in their operational work. In the last decades, a transition has been observed from dedicated trainers to the adoption of automated technologies to support the trainees. This paradigm shift makes conferring precise knowledge to novice users a challenging problem, which becomes especially relevant when the user is dealing with large and complex datasets from which to extract relevant information. In this context, the ANITA project aims at providing law enforcement agencies (LEAs) with a set of automated tools and systems to boost the investigative work in the fight against illicit trafficking activities. Within this project, one of the functionalities that have been developed is the capability to provide adequate solutions to facilitate the transfer of the acquired knowledge among users, and consequently improve the time necessary to train novel users. In particular, based on a review of state-of-the-art literature and direct feedback from LEAs, we have developed an assistive system to aid in the knowledge transfer from expert to novice officers. This system is grounded on the most relevant instructional principles derived from cognitive and learning theories. The end result is a system that can dynamically deliver suggestions based on both previous successful actions from other users and on the current performance and states of the user. To validate it, we have implemented a knowledge graph exploration task, in which users can receive suggestions from the system. Here, we propose this novel knowledge transfer system by presenting the corresponding literature review together with the conceptualization of the architecture of the assistive system and the implementation of a validation task. With this work, we aim at facilitating the transfer of domain knowledge, which could have a significant impact on the training and education of law enforcement officials in and for the Digital Age.
We are accustomed to see the police and policing as becoming increasingly technology-intensive, perhaps even technology-centred, -driven and -dependent. However, it appears pure science fiction to understand them as totally immersed by technology. Such an immersion or the sheer possibility for it seems like giving up the last hope. The attributes we attach to technical beings tend to extend the gap between technical objects and our fundamental nature as human beings. However, misunderstanding technical objects and their coming into being as well as failing to fully grasp the complexities that characterise our relation to our tools, machines, and technical ensembles have paved the way to the emergence of the prevailing understanding, where emphasis is either on freedom or alienation, mastery or slavery, the fulfilment of all human intentions or the arrival of the doomsday. I believe such conceptual shortcomings resonate with the way we see the police, especially the relation between the police and technology. As the eve for thinking machines and artificial intelligence comes ever closer, we should critically reflect the presumptions that play a constitutive role in thinking the issue. This article outlines the standard account and expands the perspective by introducing alternatives to it in the context of the police and policing.
The port of Rotterdam is an important gateway to Europe and an important logistic hub for global trade. However, factors that ensure the competitive position of the port of Rotterdam are also attractive for drug criminals. Stakeholders are becoming aware that in order to be competitive, ports must be a secure place. This awareness has led to a growing focus on security and resilience, in which digital technologies play an important role. In this paper the findings of an empirical study on the potential of AI and data science in securing ports against undermining crimes are presented . Furthermore, the requirements to realise this potential are discussed by elaborating on acceptance of new automated technologies using the Technology Acceptance Model (TAM), stressing the need for a systems approach and exploring new vulnerabilities that might arise in the highly digitalised ports in the future. The study consisted of a qualitative research, which was conducted through semi-structured interviews, in-depth interviews and an expert meeting.
The findings of this research show that developments in Data Science and AI at ports could have a strong effect on reducing the vulnerability of ports against illegal activity. In particular, it addresses the development of smart containers, which could enable better control of container movements and better protection of the container against criminal exploitation – such as the import and export of drugs. With the advent of smart technologies, the vulnerable human factor (in the context of undermining crime) in port processes could, gradually, become less important and be replaced by technology. However, new vulnerabilities may arise in the field of data ownership and cybersecurity. To realise the potential of AI and Data Science to protect ports from undermining crime, attention must be paid to these vulnerabilities, as well as ensuring the acceptance of the new (automated) technologies and adopting a systems approach.
Taking stock of what the INTERPOL Innovation Centre has learnt since its establishment 5 years ago, Ms. Hazenberg will share observations on where we are now as law enforcement. She will then present findings from high level engagements with INTERPOL member countries, including the 2nd INTERPOL Innovation Technology Advisory Group meeting and the 2nd edition of the INTERPOL Young Global Police Leaders Programme, about the future challenges confronting law enforcement – and notably about trust in a digital age. Building on these observations, Ms. Hazenberg will then give an overview of the Future of Policing, a major initiative launched by the INTERPOL Innovation Centre this year. The presentation will highlight key findings so far, in terms of the changes in the operational environment of law enforcement, what this means for the policing function and how we can prepare for these transformations, notably with regards to skills, training and development. Emphasizing the role of innovation and foresight for future preparedness, Ms. Hazenberg will then discuss outcomes from a recent foresight exercise conducted by the Innovation Centre on digital learning and training, including technological advances and practices that could enhance the experience of trainers and learners alike and better prepare police officers for the future
This paper addresses the varying levels of training preparedness and legal challenges facing the American local law enforcement agencies in the Digital Age.
From the example of the New York City Police Departments’ multiple units like: the SMART Unit (Social Media and Research Team), Real Crime Unit, Domain Awareness and Vehicle Recognition Unit to the overview of the majority of smaller police departments that have very limited, if any, type of preparedness.
The majority of police departments in the United States are staffed with less than 50 sworn officers and the Digital Age policing challenges are numerous and addressed in a very uneven manner. However, the larger departments, like the N.Y.P.D., can provide a template for a more professional and effective response.
Finally, what will be analyzed and addressed, in addition the different modalities of numerous tactical responses embedded in the creation of the specialized units, are the legal aspects of these initiatives. Some of the legal challenges facing the specialized unit will be discussed while, focusing on the hurdles in obtaining legal subpoenas for the information posted on various social media platforms like the Instagram, Facebook and Snapchat.
A round of experts will discuss the most important developments and obstacles for learning and education in the digital presence and future.
In 2016 Europol warned that: 'the global development of online infrastructures has made the Internet a crucial tool for human traffickers, and it is likely to become more significant in the future’. In 2020 Europol stated that human trafficking transformed into: 'a new business model, in which the online component is an essential part of criminals' modus operandi’. Criminals use the internet as a way to select victims and avoid detection. They do so on three levels, knowingly: I. the surfaceweb, II. the deepweb and III. the darkweb. The internet is ever spreading and new apps, sites and platforms for communication appear every day, so does the hunting ground for criminals. Criminals perceive the internet as a tool to new sources of victims and earnings with limited threats while law enforcement perceives the internet as a threat and a source of crime with limited tools to fight these crimes.
This can and needs to be turned around. The internet should not be seen as a threat but rather as an opportunity and as a way forward for both criminal investigations and intelligence as it is a way to identify online victimization and perpetration of all sorts of crimes committed online. Not only cybercrime but also cyber enabled crimes like human trafficking. The surfaceweb is the way in which victims are solicited and clients for the services are found while the deepweb is being used as a means for facilitation. The darkweb is the place where business is conducted on either online slave marketplaces or for criminals to get in contact with one and other. This however is poorly researched, insufficiently mapped out and therefor almost invisible.
The way forward would be to apply cyber ethnography which is a method of data collection that involves applying the techniques of classic anthropology and ethnography to the online world. The overall aim of cyber-ethnographic studies is to immerse oneself in the virtual world that the participants have created, in order to understand how they experience social interaction and devise ways of regulating social order.
Co-author: Sander de Koijer
From the first moment of the spread of the COVID-19 coronavirus pandemic, another pandemic broke out at the same time, that of misinformation. Social media plays a key role in this, through which any information is rapidly transmitted around the world.The consequences of transmitting invalid information can be worse than the consequences of the pandemic itself. Conspiracy theories and false therapies are just two of the common categories that have unintended consequences for public health. On the other hand, the volume of information circulated on a daily basis makes checking the reliability of information a particularly demanding challenge for law enforcement in the Digital Age.
In my Thesis, the automatic detection of fake news related to the evolving coronavirus pandemic on social networks and specifically on Twitter is studied.For this purpose, algorithms of natural language processing (NLP) and Machine Learning are utilized. The data used to train the algorithms originates from a publicly accessible dataset that contains tweets related to the current pandemic. From the dataset, only the content concerning the Greek language was isolated.These tweets were classified and annotated in three categories, true, irrelevant, or false. Once a sufficient number of data has been annotated, the most common words are visualized through wordclouds for each category. In addition, a set of linguistic and morphological features are extracted from them by applying methods of converting texts into vectors, as well as features related to the subjectivity of the tweets’ texts.Additional features are calculated using the TF-IDF method which are used in conjunction with the morphological features. Python libraries such as NLTK, spaCy and Scikit-Learn are used to calculate these features.Before these features are feeded into the learning algorithms, PCA is applied to reduce their dimensions. Three learning algorithms are trained, Random Forest, SVM and Multinomial Naive Bayes, of which Random Forest has the most encouraging results.Our results prove that it is possible to automatically detect invalid information in posts on Twitter despite the peculiarities that characterize the Greek language.
The THOR methodology is the most essential building block in a chain that aspires to influence the development of security solutions and produce meaningful recommendations to policy and decision makers. Initially, in order to identify operational gaps that security practitioners face, specific working scenarios are formulated and developed, exposing an array of possible threats and responses. Then, analysing these scenarios with different categories and ranks of security practitioners offers a well-rounded, comprehensive insight into the needs that should be fulfilled and the operational capabilities to be developed or complemented. This detailed detection and documentation of capabilities serves as an impact analysis that offers the canvas that will be used to pinpoint and categorise individual attributes that touch upon the THOR methodology’s four dimensions: the Technological, the Human, the Organisational, and the Regulatory dimension. Based on the produced findings, a strategy shall be outlined so that practitioners are able to weigh-in on the cross-over between gaps and urgencies and prioritise the fulfillment of their needs by utilising three time horizons: short-term, mid-term, and long-term.
A crucial element of the THOR methodology -and the purpose of this paper – is to demonstrate how each capability gap interacts with the THOR dimensions, revealing the interconnection of deficits. Understanding the ad hoc interplay of the four THOR dimensions is crucial to optimally grasp the challenges that need to be overcome. For instance, there are attributes that appear at first sight technological, nevertheless the sole adoption of a pertinent technological solution would not address core issues and deficits, if professional development (human-related dimension) or acquisition of expertise (organisational-related dimension), and/or a supportive legal framework (regulatory dimension) are not in place beforehand.
THOR methodology assist practitioners to carefully identify their capability needs, prioritise them, by utilizing their operational experience. The application of a multiple-dimensional approach in a field as vast as security, considers expertise and experience by various security stakeholders - their positions and specialties notwithstanding. This, in turn, generates ideas and solutions of practical value, aimed at addressing existing and emerging threats alike.
All types of crimes are deplorable and should be combated, however some of them generate more rejection in the society because of their nature. Among these are “hate crimes”, which are based in prejudices. However, societies have started to criminally punish hate crimes just in the last decades. These types of crimes involve a variety of conducts, which entails the complicated task of finding appropriate ways to prevent and tackle them. With the progress of Artificial Intelligence (AI), and more concretely NLP (Natural Language Processing), we are able to increase the efficiency of the Law Enforcement Agencies (LEA) to combat hate crimes. Here, we present two projects in which the Spanish National Office Against Hate Crimes (Oficina Nacional de Lucha Contra los Delitos de Odio, ONDOD), of the Secretariat of State for Security (Ministry of the Interior, Spain), has participated. First of all, we have collaborated in a European project called ALRECO (Hate speech, racism and xenophobia: alert and coordinated response mechanisms), which has created some algorithms to automatically detect hate speech on Twitter. Now, we are collaborating in another European project called REAL-up (Hate Speech, Racism and Xenophobia: Alert and Response Mechanisms, Upstander discourse analysis). In this new project we are trying to include other social networks despite Twitter and also counter-narrative to tackle hate speech. The algorithms, based on transformer-based models such as BERT and BETO, achieved to classify well offensive tweets in Spanish language. The dataset used, which contain more than 5000 tweets labeled, has been also created in the context of the project. This project, coordinated by the Spanish Observatory on Racism and Xenophobia (OBERAXE), it is a good example of collaboration of different Ministries, the University and the third sector. On the other hand, ONDOD is developing, in collaboration with researchers of different Universities, an automatic tool to help police officers to record properly hate crimes, so they do not misclassify those crimes in which may be doubts whether there is a prejudice.
Paper co-authored by Carlos J. Máñez, Tomás Fernández Villazala, Javier López Gutiérrez
Cultural property is a crucial part of the identity in any state and a sector with significant economic value. So, protecting it from criminals, seeking economic benefits, or any attack in which the objective would be the people's own identity (e.g. armed conflicts or terrorist attacks), is an essential duty for law enforcement agencies.
There is a vast dark market focused on the antiquities trade. If we compare it with other illicit businesses, this black market's main character is the mixing capacity with the legal market. This leads to a significant increase in the difficulties during investigations.
We must also add to this the fact that digital tools have changed the way of life and the way of businesses, and people undertaking criminal activities have not been left out of this.
In this regard, law enforcement agencies need to develop scientific knowledge and IT capacities in cooperation with academics and society to face the continuous challenges in this area.
OSINT techniques are one of the most valuable tools in this regard, such as carrying out provenance investigations, which are crucial to identifying and proving any object's illicit origin.
This presentation aims to show a quick view of the issue to foster the development of new academic research and investigations in the area.
In the modern world, the use of ICT communication technologies has become an integral part of life. ICT infrastructure is the bearer of digital traces of both legal and illegal activities that are performed through it. However, for something to become digital evidence, it must be obtained by law and by a person authorized by law to obtain it. Namely, the virtual infrastructure, especially the Internet and the new challenges brought to us by cloud architecture, due to its physical positioning outside national borders, calls into question the legality of searching and collecting digital evidence outside national borders. This presentation analyzes the legal basis for collecting digital evidence in cyberspace internationally, such as the Council of Europe Convention on Cybercrime, the US Cloud Act, the Australian Decryption Act and the European GDPR. Although the Court of Justice of the European Union declared invalid the decision of the European Commission (EU) 2016/1250 on the adequacy of data protection provided through the EU-US Privacy Shield, experts must not stop looking for a solution to the obvious problem. The intention of the presentation is to support decision makers in taking clear national positions regarding the above controversial legal norms and their mutual conflict. The presentation compares the legal consequences of such collection and the acceptability of such digital evidence, and such collection may also be associated with a breach of the privacy of a legal and private entity.
The technological developments of the last decades have granted citizens worldwide access to the Internet, including in handheld devices. Consequently, anyone virtually anywhere can post, share or comment on anything at any time. This possibility, which has given citizens freedom of choice, has also exposed them to more cybernetic attacks, while for criminals new opportunities for cybercriminal practice emerged. The latter practice has gained special visibility recently, due to the covid-19 pandemic, which demanded a world on lockdown to be connected technologically. As a result of the perceived increase of cyberattacks, Law Enforcement is faced with two major challenges: firstly, the higher the volume of cyberattacks, the harder it is for Law Enforcement to dedicate the necessary resources, including human, to fight them; secondly, and importantly, the range of sophisticated stealth technologies that can be used by cybercriminals to help them remain anonymous online seriously hamper the work of Law Enforcement. This paper builds upon research presented at the CEPOL 2017 Research and Science Conference to claim that, despite their anonymisation efforts, in a significant proportion of crimes cybercriminals use language to communicate, and since that use of language is idiosyncratic (i.e., since each speaker of language makes a particular use of the language that they speak and write (Coulthard 2004)), cybercriminals can be positively identified by the language that they use. This identification is made possible by Forensic Linguistics, which can be broadly defined as the application of linguistic analyses in legal or Law Enforcement contexts. This research presents two illustrative cases of cybercrime to show the potential of the forensic linguistic analysis. The first is the case of an anonymous set of text messages spreading defamatory contents, whose linguistic analysis enabled the sociolinguistic profiling of the author, and hence narrow down the pool of suspects. The second presents a cross border cybercriminal practice: fraudulent and deceptive messages sent to citizens for purposes of extortion. The potential of the linguistic analyses conducted in these cases, as well as other applications to uncover cybercriminal activities, will be discussed. The presentation concludes by making recommendations for Law Enforcement.
The emergence of Generative adversarial networks (GANs) has created new criminal opportunities through the generation of fake images, videos or texts, but also new possibilities to fight against crime and discrimination. This is what we call the double face of the use of GANs : Dr Jekyll and Mr Hyde. This study presents the two faces of the GAN's applications. A GAN is a generative model is based on two neural networks models : a generator and a discriminator and the combination is used to create new pausible examples. The two models are set up in a contest where the first one (generator) seeks to fool the second one (discriminator) model, and in the same time the discriminator is provided with both examples of real and generated samples. At the end, the generator is able to create new models on demand. In one hand GANs appears as a major breakthrough in the field of forgery and the potential for bad is there. It becomes easy to generate false news articles flooded almost all social media platforms, and develop a real propaganda. So, GAN's and their using in the case of deepfakes appears as a great challenge for LEA to fight against complotism, propaganda or cyber attacks on the balance of database. In another way GANs can provide relevant methods to authenticate documents, to generate photographs of faces with different ages from younger to older, in the case of child missing for instance, or to re-balance database in order to fight against discrimination and so on. Our work is on different cases using GANs and trying to understand the different faces of this new technique. We have for instance decomposed the principe of deepfake videos construction, developed a database to evaluate and to compensate bias to fight against discrimination and built a system able to detect fake images based on many GAN's methods.
Desinformation and related online threats (eg. hate speech, hacking, hybrid threats) is powerful tool to influence, manipulate and rule society. Both technological tools and human factors are important in countering it. Two Lithuanian universities – LKA and MRU – develop National Science and Studies Platform for the Recognition and Analysis of the Information Security and Information Threats (NAAS), which address above mentioned factors. The newly developed NAAS platform will integrate technological, software and methodological instruments and shall be used for training and research. Additionally, NAAS will contribute to more effective mitigation of the impact of hybrid threats at the national and international levels by ensuring public security. Not only the NAAS platform is innovative but also the method (pre commercial procurement) of developing it. Thus makes NAAS unique and much expected by security practitioners.
This paper presents policy observations and recommendations on the basis of the research findings of legal and ethical nature performed by the teams of three EU-funded projects, namely LOCARD, ROXANNE, and FORMOBILE. The main focus of the paper is to examine how cross-agency and cross-border cooperation and coordination among public and private entities with law enforcement authorities in the digital age is being impacted by the respective legal provisions pertaining to the collection, analysis, and sharing of digital evidence. In particular, we review how the right to privacy and personal data protection prompt e-evidence examination and its value in the context of the fight against crime and terrorism. The challenges posed by the different approaches applicable across the Member States with regard to the lawful access, processing and sharing of personal information is discussed, as well as potential implications to the right to defence triggered by these differences. The main challenge identified in this direction is the lack of legal regime applicable to the e-evidence handling within the criminal procedure, which further intensifies the difficulty of the practical organisation of cross-agency and cross-border pending investigations. The paper casts a look on both the strong and weak points of the current framework when it comes to the cooperation in criminal matters, namely the existing mutual legal assistance treaties and the significance of the European Investigation Order with regards to these matters. On this backdrop, the current e-Evidence proposal is critically examined, offering evidence-based suggestions on how the draft provisions could be enhanced to suit the challenges met in a cross-agency and cross-border cooperation context better.
The major investigation challenges are summarized as multiple-identity, fraudulent actions, lack of interoperability, and absence of an effective technical solution for exchanging Cross-Border information.
The EU published Regulations (EU) 2019/817 and 2019/818 for establishing a framework for EU interoperability between information systems in the field of borders and visa information systems, police and judicial cooperation, asylum, and migration. Existing systems such as EURODAC, SIS / SISII, and VIS must share data, and new systems such as ECRIS-TCN, EES, and ETIAS also need to follow these guidelines. The EU interoperability components include the European Search Portal (ESP), in addition to Europol and Interpol data; the Shared Biometric Matching Service (sBMS); the Common Identity Repository (CIR), and the Multiple Identity Detector (MID). Although the eu-LISA will implement the interoperability framework in 2023, new challenges will emerge, such as investigating multiple-identity and identity frauds due to the different formats and structures of data, low quality of biographic and biometric data, and low accuracy of matching algorithms.
Furthermore, the recent global threats such as the increase of illegal immigration, the high risks of terrorism and serious crime, the COVID-19 pandemic, and the war between Russia and Ukraine created the essential need for exchanging Cross-Border information for preventing, detecting, and investigating terrorism and serious crime across Europe and the neighboring countries.
Finally, the Open Source Intelligence (OSINT) investigation process is not automated, consumes a lot of time, and is overwhelming. When the border security and the law enforcement officers use methods of OSINT to investigate terrorism and serious crime, it is very difficult to match and link the identity-related data and facial images of the suspects stored in the EU systems, Cross-Border systems, and open sources.
The paper argues different Artificial Intelligence (AI) methods and algorithms and interoperability could be the optimum solution for the challenges mentioned above. The paper highlights a Person-Centric approach using Artificial Intelligence and interoperability to solve the challenges that emerge during investigations, such as multiple-identity, identity frauds, the complexity of OSINT investigations, and exchanging Cross-Border information.
AI hasn’t been used much yet in cybercrime, but its use is growing and experts have warned that it will transform cybercrime. The cost of cybercrime is already escalating. AI is likely to ratchet up the effectiveness and impact of cybercrimes, including cyber attacks. One of those impacts is the low rate of apprehension and prosecution of cybercriminals, less than one per cent according to some estimates. Meanwhile, the socio-economic impact of cybercrime is extensive. Cybercrime affects virtually everyone with a bank card or a mobile phone or any presence online. The EC and others have said that an improved reporting of cybercrime is desirable, as it’s widely believed that cybercrime is under reported. There are various reasons why companies are not reporting cyber attacks and not sharing more information. Companies may fear damage to their reputation or a hit to their share price or they don’t know to whom they should report a cybercrime or with whom they could usefully share information without giving a competitor an advantage. Or they just think LEAs can’t or aren’t able to do anything about it because they don’t have the resources or the necessary cybersecurity competencies, especially involving AI. CC-DRIVER and CYBERSPACE are two projects addressing these challenges. Both are in a cluster of EU-funded security projects, all of which have LEAs as partners and most of which involve AI. CC-DRIVER started the cluster, initially with eight projects in total. Since then, the number of projects in the cluster has climbed to 19. The projects share a set of objectives, an important aim of which is to leverage the impact of each project in the cluster, so far, mainly through webinars and quarterly meetings of the coordinators.
The European Anti-Cybercrime Technology Development Association (EACTDA) is a non-profit association, composed of LEAs, research&academia, and industry, that focuses on the uptake of technological solutions so that they can be offered with no license cost and access to the source code to EU LEAs and other related EU public security organisations.
Over the past few years it has been noted that decisions in the scope of public policy, including those focused on citizens’ security as well as the prevention of crime, should be based on scientific evidence. All of that with the objective of avoiding a decision making process substantiated by biased data. One of the security fields which generate more public concern is the missing persons phenomenon. In Spain, since 2018, the National Centre for Missing Persons (CNDES), dependent from the Spanish Ministry of Interior, has been promoting research in collaboration with Academia to improve instruments and investigation tools for missing person cases. Specifically, a predictive risk assessment system (“SER – DesVi”) has been developed to identify those missing person cases which result in harm, suicide or homicide outcomes. All of that is based on a checklist formed by empirical risk and protective factors which will also serve as a standardized tool for collecting historical data. A proportionate stratified sampling was conducted. In this manner, a representative sample of missing person cases was collected. For the construction of the predictive scales for harm, suicide and homicide, a relational analytical study of cases and controls was designed and bivariate (Chi-Squared) and multivariate (Binary Logistic Regression) statistical techniques were employed. Having been constructed, the scales were evaluated for validity and predictive capability using the test discrimination estimators (odd ratios – OR, area under the curve –AUC, sensitivity and specificity) and calibration estimators (positive predictive value – PPV and negative predictive value – NPV). Finally, different machine learning algorithms (Logic Rules, Decision Trees, Random Forest and Bayesian Networks) were explored to facilitate the automatic calibration of the indicators which compose the predictive scales. In the study of missing persons, the development of algorithmic and predictive models has enabled the construction and calibration of the “SER – DesVi” system which entail implications for a) the prioritisation of investigation lines, b) the adoption of a more effective strategy for the management of police investigation resources, c) the investigation of cold cases and d) the establishment of an evidence-based approach in the criminal justice system.
Complementary Coffee and Refreshment
The internet and social networks are a “new” environment for organized crime, radicalization, recruitment and disinformation. There is a deterritorialization of threats and risks, making the virtual world a new dimension for the expansion of ilegal activities and for the action of the police and justice.
Technological development has created an illusion within Police and intelligence communities of giving priority to the technical intelligence component (TECHINT), based on artificial intelligence, analytical software, big data, predictive techniques based on algorithms and many other ways of acquiring knowledge. This resulted in a gradual devaluation of community policing, human intelligence (HUMINT) and the understanding of culture, language, religion and community problems.
We propose to analyze the advantages of the Police promoting a systemic link between HUMINT and TECHINT to allow a better understanding of socio-political, cultural and religious idiosyncrasies of local communities, as well as to identify threats, risks, signs of radicalization, recruitment, organized crime and their modi operandi. The global trend of disinvestment in human sources and the prioritization in technological solutions can create an aseptic perspective of reality, the inability to detect underground criminal phenomena and increase biases in police and intelligence analyses.
We propose to present the added value of community policing, favoring the collection of intelligence through contact with human sources (citizens) and the training of the Police (namely, first responders, spotters and police analysts) evidence and intelligence gathering in local communities, as well as the cooperation between generalist police resources and specialists (criminal investigators, special operations, public order, rapid intervention, inactivation of explosives, negotiators).
We seek to prove the advantages of a synthesis between soft policing and hard policing, between community policing and intelligence-led policing, and the benefits of direct contact and dialogue between police and citizens and the development of scientific criminal analysis.
Democratic societies in a digital age need close contact between justice and the police and their citizens, the protection of individual rights and a balance between community based policing strategies and high tech and scientific techniques to detect threats and risks to our security.
The paper deals with the challenges of strategic preparedness of the police in the era of digital policing and rapidly changing security environment. The contemporary models like Intelligence-led policing, Evidence-based policing and Predictive policing rely on the quality of intelligence and information, but there is not much strategic level discussion about what is to be done with disinformation and hybrid threats. There is not much discussion, either about how to harness the power of AI and for instance LFR (Live Facial Recognition) while avoiding unethical practice and the breaching of public trust. The police is accountable also for algorithmically informed decisions. Strategic preparedness is important for public order policing and crime prevention, because the consequences of disinformation and hybrid operations are local. They can cause fear, public order and street peace disturbances, violence, targeting, demonstrations et c. Disinformation and hybrid threats are seen mainly as national security issues, which means they are political by nature and the actor may be a state, digital network or unknown. We will look at the challenges of AI in the context of new public order, which is more about resistance, refusal, activism and politics than before. The contingent nature of security and order should be taken account in development of strategic preparedness. The research methodology is based on conceptual mapping, document analysis and key expert police interviews.
Complimentary Light Lunch
In a recent EU publication, a report commissioned by the European Union related to the Cross-border Digital Criminal Justice environment, a set of specific business needs have been identified. Some of the most relevant ones have been:
The interoperability across different systems needs to be ensured.
The stakeholders need to easily manage the data and ensure its quality, allowing them to properly make use of it (e.g. use the data as evidence in a given case).
The stakeholders investigating a given case should be able to identify links between cross-border cases. Therefore, solutions are needed to allow the stakeholder to search and find relevant information they need for the case they are handling.
The study presents a set of solutions to address the highlighted needs, including:
- Judicial Cases Cross-Check (Evidence standard representation is suitable)
A Judicial Cases Cross-Check system should provide a tool being able to search for case-related information and identify links among cases that are being investigated in other Member States or by JHA agencies and EU bodies.
To facilitate the development of the above solution, a standard representation of the metadata and data of the Evidence should be adopted. In particular the ontology UCO/CASE, dedicated to the digital forensic domain, seems the most promising one to this aim. UCO/CASE, that stands for Unified Cyber Ontology / Cyber-investigation Analysis Standard Expression, provides a structured specification for representing information that are analysed and exchanged during investigations involving digital evidence. To perform digital investigations effectively, there is a pressing need to harmonise how information relevant to cyber-investigations is represented and exchanged. CASE enables the merge of information from different data sources and forensic tool outputs to allow more comprehensive and cohesive analysis. All these metadata represented in a standard format, could be provided to any potential stakeholder using a decentralised repository of metadata along with a suitable level of confidentiality and integrity.
The INSPECTr project (inspectr-project.eu) opted for the open-source UCO/CASE ontology to serve as a standard for interchange, interoperability, and analysis of investigative information.
The datafication throughout different spheres of our societies has led to an increasing reliance on data, even more so with the advances of artificial intelligence (AI) methods. A development that can also be observed in the sphere of law enforcement and crime analysis. Tools like predictive policing, that rely on the quantitative analysis of past crime data, are used to inform police operations. Ideally – as is indicative of the name – by predicting future criminal behaviour and occurrences. But also, other AI supported tools are increasingly used in a policing context: facial recognition, shot spotters, crime scene analysis tools; to name just a few. All these tools have in common that they are built with the purpose to make police work and crime fighting more efficient, improving the routines and practices of policing, and lead to faster and more pre-emptive insights.
That being said, the use of these AI-tools doesn’t come without a caveat. There are a broad range of examples that show the unintended side-effects of AI use in policing. These range from a shift of financial investment into intelligence led policing at the expense of other areas of law enforcement, to highly biased and discriminatory decisions in policing. The first brings forth problems within the police as an organisation, the latter risks solidifying discriminatory practices through the materialization of AI-technologies. In this paper, I will provide an analysis of these problems, which have similarly emerged in other areas of society – such as (health) care, or consumer research. While I will first expand on how these problems emerge and try to indicate their causes, I will discuss in a second step some ideas on how these issues can be acknowledge and resolved – at least partially.
Mobile devices have become an indispensable part of modern society and are used throughout the world on a daily basis. The proliferation of such devices has rendered them as a crucial part of criminal investigations and has led to the rapid advancement of the scientific field of Mobile Forensics. The forensic examination of mobile devices provides essential information for authorities in the investigation of cases and their relative importance advances as more evidence and traces of criminal activity can be acquired through the analysis of the corresponding forensic artifacts. Big Data forensics, in particular, requires a new generation of technologies and architectures, designed to efficiently extract value from very large volumes of a wide variety of data, which can significantly contribute, with correct technical interpretation and correlation through expert analysis, to the successful completion of criminal investigations. In the age of Information, the need for advanced forensic tools that will utilize the most prominent achievements in Data Science has become critical. We will examine the current status of Mobile Forensics in relation to the field of Big Data, by exploring the most important challenges and investigating the existing Artificial Intelligence and Machine Learning solutions. The utilization of these emerging technologies provides crucial tools and enhances the professional expertise of law enforcement and forensic scientists, paving the way to overcome the critical challenges of criminal investigation.
In 2021 Frontex conducted a Technology Foresight on Biometrics for the Future of Travel, with the objective of studying the future of biometrics for its implementation in border check systems that may benefit the work of the European Border and Coast Guard community in the short-, medium- and long-term perspectives.
Three experts’ consultation events (two Technology Foresight Workshops and a Delphi survey) took place during the project. A broad group of relevant stakeholders was involved in these events to exploit collective intelligence and stimulate consensus-oriented discussions.
A custom Technology Foresight methodology was developed, opening the door to the exploration of the vast field of biometric technologies, which were analysed from various perspectives in the context of border checks.
Each of the phases of this complex Research Study produced its own set of insights which include:
• a thorough explanation of the overarching Technology Foresight methodology and tools adopted, customised to the needs of the study but with future implementations in mind;
• a taxonomy of biometric technologies and biometrics-enabled technological Systems;
• the results of the in-depth analyses conducted on patents, scientific literature and EU-funded projects;
• a customised set of scenarios for EU in 2040;
• a prioritisation matrix of biometric technological clusters;
• a set of roadmaps developed for the identified key technological clusters (KTCs - contactless friction ridge recognition, 3D and infrared face recognition, iris recognition in the near-infrared and visible spectrum);
• a set of capability readiness heatmaps showing an overview of the extent to which KTCs’ cluster-specific needs are met at present or will be fulfilled in the future under the different hypothetical scenarios.
Due to the substantial amount of information provided and the adopted participatory foresight approach, the Research Study will directly contribute to an enhanced understanding of the relevance and applicability of novel biometrics and technology foresight. The findings can also be used by public organisations, research and technology organisations, academia and industrial entities in Europe to identify areas of strategic interest and to make informed decisions about paths of future developments in biometrics, acting towards strengthening European strategic autonomy in this sector.
The purpose of this paper is to explore the development of an online law enforcement organisation. During the pandemic, many international projects had to find new ways to implement their projects very quickly. In international cooperation, work transferred to an online environment. This required flexibility and new ways of working for organisations and experts involved. Development work also faced new challenges. This study explored the challenges faced by experts and the innovative solutions they have used. At the same time, it explores the challenges encountered in the organisations that have been involved in the development process. By comparing the data, it will be possible to find out whether the experts involved in the development work experience similar challenges as the staff of the organisation being developed. The experiences of Estonian and Finnish experts will also be compared. At the same time, the aim is to identify good practices that can be applied in international cooperation. The research will focus on the Twinnig project "Strengtening the Capacities and Efficiency of the Security Academy - Albania", which started in 2020. The study will be based on a survey of experts (Finland and Estonia) and interviews with key persons in the organisations targeted by the project. The organisations, trainers and experts involved in international development work will benefit from the results of the survey. The good practices reported in the study can also be used in international online training.
The use by criminals of complex cross-border corporate structures represents one of the greatest challenges for financial crime investigators. This study has developed a predictive model which improves the early detection of high-risk firms and owners through the identification of red-flags and anomalies in firms’ ownership structures. To this end, we collect and analyse business ownership data on a sample of more than 3 million firms registered in nine European countries. To validate the risk indicators, we implement several machine learning models that are trained and tested against evidence of sanctions and enforcement. The proposed risk models have a strong predictive power. Firms with (i) more complex structures, (ii) owners from high-risk jurisdictions and (iii) links to opaque corporate vehicles are more likely to engage in illicit activities. The inclusion of macrolevel information, such as geographic location and business sector, significantly improves the understanding of the phenomenon. The risk assessment model is then incorporated into a novel prototype tool which would allow supervisory and investigative authorities to real-time screen and investigate large samples of firms, allowing them to prioritise and speed-up investigations. This paper is partially based on the results of EU co-funded project DATACROS.
Paper co-authored with Maria Jofre and Michele Riccardi.
The INSPECTr project aims to produce a proof of concept that will demonstrate solutions to many of the issues faced by institutional procedures within law enforcement agencies (LEAs) for combating cybercrime. Unlike other H2020 projects, the results of INSPECTr will be freely available to stakeholders at the end of the project, despite having a low technology readiness level. It is imperative that LEAs fully understand the legal, security and ethical requirements for using disruptive and advanced technologies, particularly with a platform that will provide AI assisted decision making, facilitate intelligence gathering from online data sources and redefine how evidential data is discovered in other jurisdictions and exchanged. However INSPECTr will also require the support of stakeholders beyond the scope of the project, in order to drive further development and investment towards market-readiness. The development of a robust capacity building program has been included in the project to ensure that LEAs can confidently use the system and that they fully understand both the pitfalls and the potential of the platform.
During our training needs analyses, various European instruments, standards and priorities are considered, such as CEPOL’s EU Strategic Training Needs Assessment, the course development standards established by ECTEG and Europol’s Training Competency Framework. With this research and through consultation with internal and external stakeholders, we define the pathways of training for the INSPECTr platform in which we aim to address the various roles in European LEAs and their requirements for the effective delivery and assessment of the course. In keeping with the project’s ethics-by-design approach, the training program produced by INSPECTr will have a strong emphasis on security and the fundamental rights of citizens while addressing the gaps in capabilities and training within the EU LEA community. In this paper we describe the process we apply to curriculum design, based on the findings of our research and our continued engagement with LEA and technical partners throughout the life-cycle of the project.
Law enforcement practitioners today face increasing challenges when suppressing organized crime activities. Based on the responses of more than 120 LEA practitioners around the world who participated in a global online survey that was conducted by the ROXANNE consortium in Spring 2020, key pain points were the vast amount of data to be processed for each case, which is increasingly heterogeneous, and that crime evolves and adapts to new realities becoming increasingly more complex.
The high velocity and vast size of collected data calls for an efficient way to collect, filter, analyse data and automate the most time-consuming tasks. ROXANNE combines multiple analytics on various modalities (e.g., audio, text) to support LEAs in their investigations, by focusing their attention on potentially relevant information hidden in large volumes of data. Key ROXANNE technologies include Speaker Identification (SID), Natural Language Processing (NLP), Video analysis (VA) and criminal Network Analysis (NA). Nevertheless, according to the same study above, 90% of the participants have not received any training at all around one or more of the technologies above, while their level of knowledge varies mostly from no knowledge at all to basic knowledge.
Furthermore, ROXANNE offers an intuitive user interface supporting (criminal) Network Analysis for understanding how entities interact, Timeline Analysis for uncovering patterns of interactions on the temporal dimension, Statistical Analysis for viewing data distributions and Advanced Filtering for narrowing down the results according to users’ criteria. Nevertheless, machine-driven analysis is not a silver bullet and its limitations (e.g., non-zero false positive rates) mandate human expertise and judgement to complete the investigation. Thus, ROXANNE shall allow practitioners to judge and update the outputs, improve the accuracy so that the AI-based technologies are improved in the long run.
In the proposed session we plan to give an overview of ROXANNE platform and demonstrate it using a synthetic dataset for a fictional drug-dealing case that includes more than 100 wiretapped calls, which was prepared by the ROXANNE consortium and has been used in several training sessions.
The presentation is about CYCLOPES - Network of European law enforcement practitioners fighting cybercrime. CYCLOPES is funded by EC through Horizon 2020 Programme. Within this project we identify current capabilities of LEAs as well as gaps and needs in specific areas related to fighting cybercrime. Moreover wo do the market and research scan searching for good solutions, reposing for LEAs gaps and needs. Project provide different activities for practitioners, including Join Live Exercises, were practitioners are able to test different kind of tools and solutions.
As the rules designed to counter money laundering constantly change, criminals find new methods and platforms to launder their “dirty” money. For example, such new platforms include the art market and the use of crypto currencies both of which have recently been added to the list of sectors susceptible to facilitate money laundering. Apart from the traditional art market, criminals may use digital art in order to facilitate their activities. The rise of the digital art market with the expansion of Non-Fungible Tokens (NFTs) is a new area of concern for law enforcement agencies (LEAs). Anonymity and price volatility of NFTs create a unique environment for criminals. The complex nature and uncertain legal status of NFTs further complicate the counter measures one can take. The purpose of this paper is to present NFTs, analyse their relation to money laundering and art, scrutinise their legal status in the EU and provide potential regulatory solutions and recommendations.
In addition, the paper will examine the difficulties faced by LEAs in the training and education of enforcement officials when such novel technological advancements make their appearance. The fast techonological advancements are, often, difficult to be followed by officials and academics. As training and education are a significant part of enforcement officials, the fast technological advancements create a gap that may have negative results in enforcement. As a result, multidisciplinary research from academics (both form applied sciences and social sciences) is necessary to close this knowledge gap. NFTs are such an example of quick expansion of a new technological phenomenon that highlight the urgent need for training and education needs of LEAs via public-private partnership models.
In the modern world of innovation and technology, the teaching/learning environment is becoming more modern, new didactic tools are being introduced, and the teacher-learner relationship based on equivalence is being revisited. At the same time, teaching methods are also being reconsidered. However, it is still debatable whether classical didactic teaching tools and formats are still relevant in this era of digitalization, when electronic space has become an integral part of everyday learning, and virtualized and algorithmized teaching solutions are gaining momentum.
The aim of the research is to assess whether conceptualized traditional didactic tools are appropriate for the investigation of modern crimes.
The results of the research showed that in a changing dynamic environment, where new offenses are being criminalized, the parameters of criminal offenses themselves and their assessment are changing, the divide between criminal and administrative liability is being re-evaluated, and the advancing ways of committing crimes are being considered. Therefore, the development and revision of criminalistic methodics for the investigation of certain types of criminal offences as a conceptualized format for classical teaching is a continuous process in response to the cardinal dynamism of the environment, requiring the integration/addition of up-to-date techniques and tools of investigation. The modeling of the topicalities of the investigation of one of the modern types of crimes in the light of the classical criminalistic methodics revealed that traditional didactic tools are appropriate for the investigation of modern crimes. However, when developing, reviewing, and updating criminalistic methodics, it is necessary to identify those elements of the content of the methodics that are relevant to the investigation of specific modern criminal offences taking into account the conditions of that time, new crime patterns, innovations applied in investigation, and other issues related to investigation.
Complimentary Coffee and Refreshment
The challenge for internal security practitioners including law enforcement and the justice sector is to determine how to capitalise on the opportunities offered by Artificial Intelligence (AI) and Machine Learning to improve the way investigators, prosecutors, judges or border guards carry out their mission of keeping citizens safe and rendering justice while, at the same time, safeguarding and demonstrating true accountability of AI use towards society.
The AP4AI (Accountability Principles for Artificial Intelligence) Project addresses this challenge by creating a global Framework for AI Accountability for Policing, Security and Justice. The AP4AI Framework is grounded in empirically verified Accountability Principles for AI as carefully researched and accessible standard, which supports internal security practitioners in implementing AI and Machine Learning tools in an accountable and transparent manner and in line with EU values and fundamental rights.
The project has defined and validated set of universal Accountability for AI Principles that internal security practitioners including the justice sector may adopt in order to demonstrate accountability in their use of AI. The principles are universal and jurisdiction-neutral, to be used as a guide by internal security and justice practitioners globally to support existing governance and accountability mechanisms through self-audit, monitoring and review. The AP4AI Project provides an implementation blueprint for the deployment of AI for internal security practitioners.
The AP4AI Project is jointly conducted by CENTRIC and Europol and supported by the EU Agency for Fundamental Rights (FRA), Eurojust, the EU Agency for Asylum (EUAA) and the EU Agency for Police Training (CEPOL) in the framework of the EU Innovation Hub for Internal Security.
This submission will present the project narrative, approach as well as results of the project and their relevance for the internal security domain.
Co-authors: Saskia P. Bayerl & Babak Akhgar
Digitization and algorithmic tools have over the past years found their way into law enforcement contexts, notably including the interconnection of large-scale databases, biometric identification and matching, automated surveillance capacities, short-term situational predictions, and AI-supported analysis for large amounts of data. These tools can help to increase the effectiveness and efficiency of law enforcement operations on the strategic, tactical, and operational level.
They do, however, also come with a number of concerns that must be acknowledged and addressed in order to realize their potential and avoid unintended side-effects and societal frictions. Based on a multi-year research project on predictive policing in Germany and Switzerland, this contribution will provide a systematic perspective on the challenges involved in implementing new and emerging technologies in law enforcement contexts.
Specifically, it will address (1) the nature of data, i.e. how data are socially constructed and present a particular account of the world, inevitably leading to “biased” results; (2) transparency in algorithms and AI, i.e. how “black boxes” undercut human capacities to understand and retrace processes and create problems for public accountability; (3) automation and human control, i.e. the question how human operators can retain meaningful influence over analytical processes; (4) decision-making processes and automation bias, i.e. how humans can be empowered to critically question and override system recommendations; and (5) strategic and societal implications, i.e. the fact that digital tools should not be misused to displace larger programs that address the root causes of crime.
In summary, this contribution will seek to offer a critical practical perspective on how to implement new and emerging technologies in a responsible way, paying attention to human rights and civil liberties implications as much as to potential internal friction and resistance.
Experts discuss the role and relevance of Artificial Intelligence for Law Enforcement Training and Practice
The paper analyzes several aspects of the video surveillance system application, starting from the prevention of misdemeanors and crime according to the Council Decision on the establishment of the European Crime Prevention Network. The second aspect relates to the use of video surveillance systems in the misdemeanors and crime investigation, and the third one relates to the evidential value of video surveillance systems in court proceedings. For this purpose, the case law analysis of the highest level was made, namely of the High Misdemeanor Court, the Supreme Court of the Republic of Croatia and of the European Court of Human Rights through case studies. The paper discusses the evidential value of the footage important for criminal investigation. However, the central issue is a question whether digital evidence in the form of video surveillance can be decisive in court proceedings or not, since no court order is required as it for other evidentiary actions. The paper proposes solutions de lege ferenda given that video surveillance systems are becoming more widespread and have proven to be very effective in criminal investigation, but, contextually speaking, also in procedural terms. The respective contextual approach requires the interpretation of current case law emphasizing that the content and significance of the footage in court proceedings must be perceived as a whole and that besides the right of defense, the public and the victim’s interests are to be taken into account.
Life Demonstration of Project Applications and Products
With the evolving threat of climate change and the consequences of industrial accidents to becoming more severe, there is an increasing need for First Responders to access reliable and agile information management systems that offer as higher Situational Awareness and better Common Operational Picture. To match with current trends, the RESPOND-A project aims at developing holistic and easy-to-use solutions for First Responders by bringing together the complementary strengths of its Investigators in 5G wireless communications, Augmented and Virtual Reality, autonomous robot and unmanned aerial vehicle coordination, intelligent wearable sensors and smart monitoring, geovisual analytics and immersive geospatial data analysis, passive and active localisation and tracking, and interactive multi-view 360o video streaming.
The synergy of such cutting-edge technological advancements is likely to provide high-end and continuous flows of data, voice and video information to First Responders and their Command & Control Centres for predicting and assessing the various incidents readily and reliably, and saving lives more efficiently and effectively, while maximising the safeguarding of themselves, before, during and after disasters. To this end, RESPOND-A envisions at exercising First Responders for getting familiar with the project technological outcomes, and demonstrating their real-world performance and effectiveness in the classified training facilities of our Responder Partners under hydrometeorological, geophysical and technological disaster scenarios.
On average 150 murders take place in the Netherlands yearly, but not all such incidents can be solved. Currently there are more than 1700 unsolved cases that classify as a "cold case" by the National Police Netherlands. Investigation into these types of capital offenses takes a lot of time, money and capacity. Applications of the current working method and available techniques are very labor-intensive and time-consuming. In addition, the pressure on the executive police officers is high - from the police organization, the Public Prosecution Service, the media, the next of kin, and society in general.
From an investigative point of view, it is important to be able to provide direction in the criminal investigation, in which ‘tunnel vision’ should be prevented. From a scientific point of view, more research into homicide cases in the Netherlands is of eminent importance. Remarkably little has been written in scientific literature about this type of crime.
Saxion University of Applied Sciences in The Netherlands in collaboration with the Dutch Police Academy and a number of partner companies initiated a research line to investigate to what extent data science techniques and modern digital technologies can be used to organize, direct and manage murder and manslaughter investigations more effectively and efficiently. The first project in this line – ‘Cold Case: Solved & Unsolved’ – focused on the use of open sources to collect the data and gain more insight into homicide cases in The Netherlands, explored applicability of modern data processing and analysis techniques, as well as developed a pilot tool for structuring the investigative process. A so-called ‘narrative approach’ and ‘scenario-thinking’ served as a backbone for the development of new techniques. That was also the first step in building the biggest homicide database using open-source data. The project was integrated with a study course ‘Minor Cold Case’ in Saxion University to facilitate the use of the developed knowledge and tools in the educational programs.
As a follow-up, a collaboration with a volunteer-based organization dealing with cold case research will be set up for further data collection and development of new tools.
A free, democratic and open EU provides endless opportunities for its people. However, growth is not without risk, especially in cyberspace, in the ubiquity of connected devices and rapid technological change. Criminality is also adapting, seeking opportunity and taking on new forms. While Europol’s IOCTA (2018) report described a number of positive legislative and technological developments, for example, the General Data Protection Regulation (GDPR), the Network and Information Security (NIS) directive and 5G technology, it also highlighted that “all will in some way impact on our ability as law enforcement officers to effectively investigate cybercrime”. Europol emphasised the need for law enforcement to engage with policymakers, legislators and industry to “have a voice in how our society develops.”
In this sense, CC-Driver and RAYUELA projects (which are funded by the Horizon2020 EU Framework Programme) have brought together Law Enforcement Agencies (LEAs), sociologists, psychologists, anthropologists, legal experts, computer scientists and engineers, to develop novel methodologies that allow better understanding the factors affecting online behaviour related to new ways of cyber criminality.
CC-DRIVER and RAYUELA projects are using a multidisciplinary approach to investigate, identify, understand and explain drivers of new forms of criminality, with Valencia Local Police acting as Liaison Officer. These projects focus on human factors that determine criminal behaviours such as cyber juvenile delinquency and adolescent hacking. Scientific investigation of drivers into cybercrime, impact of online disinhibition and the effect of youth decision-making processes will inform our evidence-based intervention, mitigation and deterrence strategies.
With the technological development and the increased access of people of different ages to more and more devices connected to the Internet, crime has either shifted from the real life environment to the online world, either new vulnerabilities and types of attacks have emerged that can only take place online. Without proper knowledge and appropriate skills, the risk of becoming a victim of cybercrime is high. Children are surrounded and have instant access to a myriad of information and if they are not taught how to handle it, they can easily endanger themselves or others. Therefore, law enforcement has to tackle cyber offences in which children are involved more often than before and must be kept up to date to the latest technologies and special juvenile hearing techniques to investigate these types of cases. This paper, through a qualitative analysis of interviews with law enforcement officers and with teachers from ten counties in Romania, intends to identify the main challenges that the people dealing with children have faced in handling cyber cases and ways in which the investigation and knowledge about such crimes can be improved. Furthermore, an opinion poll was conducted with children aged 10 to 18 years old to find out the unpleasant situations which they have been confronted with on the internet. The results will be used to train police officers in Romania which deal with prevention or investigation of cybercrimes and to make them more aware of the conducts that children have on the Internet, the risky situations which they are confronted with and with what determines them to commit or become a victim of such crimes, according to their age.
Digitalization transforms our homes, workplaces, and lives in general. It is thereby an established way of shaping and optimizing organizations on all managerial levels. By following the needs for digitalization, the public sector faces similar difficulties as the private sector, even though law enforcement agencies do not create a monetary value in digitalizing their processes with the equal purpose as private institutions.
Regarding this background, the KIRAS-Project DIGDOK examines the possibilities for optimizing analog forms of inmate-related documentation in the Austrian penal system. In its particular research field, the project aims at the analysis of digitalization potential in prisons including gaps between organizational leadership functions and requirements, law, and the perspective of front-line end users. Even though, the variance between law enforcement and penal system, the two organizational units are inevitably linked, especially in terms of their exchange of information, public-organization specifics, ethical aspects, and inmates' rights.
Accordingly, developing an optimized end-product is a balancing act between all various stakeholders and therefore requires an initial open field approach increasingly condensing with the framework conditions and management requirements of the public service. Therefore, we conceived a mixed-method research design using qualitative empirical methods and business informatics methods in a circular procedure of grounded theory based situational analysis to fulfill this aim. According to the needs-oriented principle, the development of successful and sustainable technological solutions depends on the satisfaction and acceptance of end users regarding the benefit for their daily work.
Finally, digitalization and the use of technology change the everyday behavior of practitioners. Thus it has an impact on their particular use of discretion, which plays a central role on daily interaction in prisons and is at least challenging the organization structure of traditionally analogue-bureaucratically structured public institutions. By observing these changing spaces of action using technology, the research findings finally allow assumptions in two areas: first, the change and its potential for optimization of workflows and second in an organizational context of action theory, the more implicit influence on practitioners’ daily decision makings.
DARLENE aims to investigate means by which AR can be deployed in real time to aid in LEA decision-making by employing AR capabilities and combining them with powerful ML algorithms, sensor information fusion techniques, 3D reconstruction, wearable technology and personalized context-aware recommendations. Hence DARLENE will offer European LEAs a proactive security solution which will provide an IoT level of Situational Awareness, detection and recognition, combining cutting edge technology and public security in all security verticals. It will enable LEAs to reduce and prevent crime, and to more quickly respond to crimes already in progress, by enabling them to sort through massive volumes of data to predict, anticipate and prevent criminal activities, make better informed tactical decisions and provide enhanced protection services for European citizens.
DARLENE will therefore develop practical and beneficial policing applications through the use of affordable, light-weight and inconspicuous AR glasses. Such applications will capitalize on cutting edge research that will combine with AR Technology to create innovative methods for combating crime and even terrorist acts. To align technology development with actual security needs and requirements, DARLENE will build a community of LEAs organization and security stakeholders that will guide the development process and evaluate the entire DARLENE ecosystem. In this regard, DARLENE foresees extensive demonstration and training activities for LEAs while the entire solution will be demonstrated and validated in realistic scenarios during the pilot phase of the project, thus paving the way to its field deployment and commercial uptake.
Live streaming has led to a relatively new phenomenon in online child abuse (Jeney, 2015).The live streaming of child sexual abuse, also known as Live Distant Child Abuse (LDCA), comprises
the instantaneous airing of a child being forced or coerced into performing or engaging in sexual acts in front of a webcam (International Justice Mission, 2020). The COVID-19 pandemic has led to a significant impact on this type of child exploitation. With offenders experiencing restrictions in physical contact opportunities, the availability of live-streaming services has served as an alternative way to accessing children flaming an increase in this category of abuse.
Not all live streaming child exploitation is carried out through pay-per-view services. Evidence suggests that certain live streaming services offer a meeting point as opposed to direct sexual performances. This allows offenders to operate beneath the radar in what appears to be a legitimate environment. Evidence indicates that offenders operate under the watchful eye of moderators and have developed techniques to work around controls. This study explores the behaviour of broadcasters and their audiences in an attempt to understand the dynamics of this category of live streaming services. Popular streaming sites such as vk.com, ok.ru, periscope.tv and twitch.com were used as backdrop to understand the relationship between young broadcasters and their audience. All sites observed appear to offer legitimate services that are not purposely designed for child exploitation. Observations of live broadcasts focused on texts posted by participants during sessions and the overall demeanour of the audience. A typology of audience interaction patterns and behaviours emerges offering us a better understanding of their efforts to grab broadcaster attention and trust. The manner in which offenders interact with potential victims differs demanding a tailor made approach towards providing an effective response. This understanding offers us a better opportunity to stop the abuse at source through the introduction of direct interventions intended to safeguard the interest of young internet users.
Cybercrime is a borderless crime that leverages technology and the internet to exploit businesses, communities and individuals. Law enforcement officers responsible for investigating cybercrime need to be equally able to access cutting edge technology to combat these crimes and to bring down criminal networks.
This talk presents the EU-funded INSPECTr project as a solution to many of the issues face by law enforcement agencies. INSPECTr produces and integrates a range of high-tech approaches, including big data analytics, cognitive machine learning and blockchain technologies into a shared intelligence platform that will improve digital investigations and forensic capabilities, and reduce the complexity and cost of cross-border collaboration. The platform is being designed through extensive collaboration with the law enforcement community, incorporates privacy and ethics by design principles, and takes into account relevant national and international legislation.
After the project, the platform will be freely available to the law enforcement community and adoption will be enhanced through training courses, webinars and workshops. Exploitation of the project deliverables will also be available to LEAs who wish to further improve the platform, beyond the scope of the project, through additional research and development activities.
Complimentary Coffee and Refreshment
As there are many different fields of application for AI driven technologies, including law enforcement, as well as differences in the scale and impact on people, using AI systems implicates a wide spectrum of fundamental rights. Therefore, it is crucial that the application of AI (and its future regulation) is firmly grounded in respect for fundamental rights. Since 2019, FRA has produced a series of reports assessing the fundamental rights implications of AI and big data, and has repeatedly underlined that only a rights-based approach guarantees a high level of protection against the possible misuse of new technologies. During the presentation, reference will be made to the findings of FRA’s published research, alongside the Agency’s on-going and future research in the field of AI - which includes the area of law enforcement.
Mr Onidi will address the conference with a brief outlook on the relevance of training for law enforcement in the digital age.
Wrap-up and farewell to participants
Complimentary Light Lunch