Automated Decision-Making Systems
The research project aims to investigate how automated processing of large volumes of personal data can support conflict of interest assessment by addressing the inherent informational asymmetry in public administration, known as "conflict of interest certification". To this end, we have studied automated decision-making systems, which involve making decisions without direct human involvement. These decisions may be based on factual data, digitally created profiles, or inferred data. On this page, various European projects are listed. Expanding each individual project allows access to detailed information such as related links, funding, total costs, principal investigator, tool functionality, project partners, and tool users.
List of Tools
Link: https://www.project-aida.eu
Funding: Horizon 2020 European Programme
Total Cost: €1,499,690
Principal Investigator: Katholieke Universiteit Leuven
Location: Belgium
ADM in action: Detect and monitor relevant data sources for investigations into cybercrime and terrorism in the Surface Web, Deep Web, and Darknet; Handle vast amounts of heterogeneous data sets (text, images, videos, geospatial intelligence, communication data, traffic data, financial transaction data, etc.) from open sources (e.g., social media platforms, blogs, forums, etc.) and privileged sources (e.g., information gathered during a national investigation of a target); Transform data (structured or unstructured) into actionable intelligence; Detect patterns related to anomalous/suspicious behaviors, suspicious events, and criminal intentions of individuals and groups that would not be visible without AI capabilities; Predict emerging trends and activities in cybercrime and terrorism, informing policy formulation as well as law enforcement; Extract and summarize multilingual conversations; Reveal content of hate speech, denigration, and misinformation; Link relevant investigative content with original authors; Discover online criminal and terrorist communities and groups; Ensure a standardized chain of custody for further forensic analysis; Enable correlations between cases.
Partners: Engineering, Europol, ItI, Centric, Vicomtech, Pluribus One, ITI, Bitdefender, Expert, System, Lingea, Voice Interaction, University College Dublin, Cybercrime Research, Institute, Kemea, PSNI, Guardia Civil, Hellenic Police, Inspectoratul General al Politiei Romane, Dutch National Police, Policia Judiciaria, Estonian Police
Funding: French Government
Principal Investigator: French Government
State: France
ADM in action: Alicem is designed to allow citizens to register their face’s biometric characteristics using a smartphone app. Once registered, citizens can use the app to complete administrative procedures that would otherwise require physical visits to government agencies. Alicem could act as a centralized database for citizens’ face biometrics, although the government claims no biometric data is stored. The app’s potential applications could include widespread face recognition via video surveillance cameras.
Link: https://aligner-h2020.eu
Funding: Horizon 2020 European Programme
Total Cost: €1,499,960.00
Principal Investigator: Fraunhofer Gesellschaft zur Förderung der Angewandten Forschung
ADM in action: Establish a series of regular workshops where stakeholders from law enforcement, civil society, policymaking, research, and industry will exchange topics related to the use of AI by law enforcement; Cover emerging crime patterns, capacity-building needs, and the ethical, legal, and social implications of AI usage by law enforcement; Support the workshops with an AI technology surveillance process and ethical and legal assessments; Publish the workshop results in the AI research and policy roadmap.
Partners: Fraunhofer, CBRNE, FOI, Ku Leuven, Gobierno Vasco, Swedish Police, Bayerisches Staatsministerium des Innern, für Sport und Integration, Germany
Funding: Belgian Government
Principal Investigator: The public employment service of Flanders, Katholieke Universiteit Leuven, Vlerick Business School
State: Belgium
ADM in action: The system analyzes thousands of job seeker profiles and examines the click behavior of individuals looking for jobs on the VDAB site. According to VDAB, this process has significant predictive value concerning long-term unemployment. The information is intended to enable early and more efficient intervention. One of VDAB's goals is to determine if click behavior analysis can be used to monitor and influence the active job search behavior of users. Job seekers who are not sufficiently active online would be invited for an interview, with the next step potentially being a penalty. Another application of this technology would be similar to Amazon's recommendation system. Based on the vast amounts of data collected by VDAB, the system could provide users with a list of recommended jobs and connect potential employers with suitable candidates. VDAB believes that using these data-driven methods can improve the personal guidance of job seekers.
Link: http://anita-project.eu/it
Funding: European Union
Total Cost: €4,999,580.00
Principal Investigator: Engineering
Location: Italy
ADM in action: Conduct advanced analysis of online data sources and cryptocurrency transactions, including the use of blockchain technologies; Utilize advanced Big Data Analytics tools to analyze large amounts of multimodal content; Develop sophisticated methodologies to capture, model, and infer knowledge, including the use of expressive ontologies and neural networks; Implement a cognitive and adaptive user modeling framework to integrate human cognitive functions into the analysis process and facilitate domain knowledge transfer; Create domain-related intelligence applications to identify patterns of correlations between illegal events and support decision-making processes for countermeasures.
Partners: Engineering, Expert System, Systran, RISSC, JADS, Italian Institute for Privacy and Data Valorization, POLITIE, DITSS, Institute for Bioengineering of Catalonia, Austrian Institute of Technology, Belgian Road Safety Institute
Tool users: EUROPOL, INTERPOL, ENFSI, ENLETS, I-LEAD, ILEANET, LEA
Link: https://www.ap4ai.eu
Funding: European Union
Principal Investigator: Europol, EU Innovation Hub, Centric
ADM in action: Develop and enhance the functionality and content of the tool in collaboration with EU AI agencies, evaluated by various European law enforcement agencies active in the field; Expand and update the tool once the AI Act is adopted, as the regulation is still under negotiation; Provide a web-based tool designed to support internal security professionals in assessing the compliance of their AI systems with the requirements of the AI Act; Enable users to verify whether existing or future applications meet the criteria set by the new regulatory framework.
Tool Users: Internal security agencies
Link: https://appraise-h2020.eu
Funding: Horizon 2020 European Programme
Total Cost: €9,425,581.75
Principal Investigator: CS Group - France
Location: France
ADM in action: Develop a scalable and efficient data intelligence platform for threat detection; Provide actionable information to proactively detect vulnerabilities and analyze ongoing crimes or terrorist attacks; Conduct soft target risk assessments based on web content, social media analysis, and on-site sensor data; Ensure instant situational awareness for planning and executing mitigation actions; Facilitate collaboration to mitigate incidents from their initial detection.
Partners: CS, Engineering, Itti, Alchera Technologies, Hololight, Aitek, Ethics, Iti, Inov, Vicomtech, Centric, ICS, Polish Platform, Local Police of Turin, Austrian Police, Slovenian Police, BTC, Links, Aòse, List, Astrial, Raid Police Nationale
Link: http://www.arno.ee/
Principal Investigator: City of Tartu
State: Estonia
ADM in action: for submitting and processing applications for nursery schools and childcare services. Tartu employs ADM systems to determine a child’s municipal school based on the address registered in the population register. The school placement offering is generated automatically once a year and is based on three sets of data: the automatic queue for a pre-school-aged child, the parents’ preferred kindergarten, and the nearest kindergarten. Parents can apply to change schools at a later stage, depending on specific circumstances. For kindergartens, parents need to apply for a spot as soon as the child is born.
Partners: OU Piksel
Link: https://www.asgard-project.eu/
Funding: Horizon 2020 European Programme
Total Cost: €11,992,533.25
Principal Investigator: Fundacion Centro de Tecnologias de Interaccion Visual y Comunicaciones Vicomtech
Location: Spain
ADM in action: ASGARD aims to contribute to LEA's technological autonomy by building a sustainable and enduring community for Law Enforcement Agencies (LEA) and the R&D industry. This community will develop, maintain, and evolve a best-of-class toolkit for the extraction, fusion, exchange, and analysis of Big Data, including cybercrime data for forensic investigations. ASGARD will help LEAs significantly enhance their analytical capabilities. With forensics as a focal point of ASGARD, both intelligence and prediction dimensions are addressed by the project.
Partners: aditess, ait, bsc, Belgian police, bmi, kemea, iti, cea, dcu, engineering, ibm, Italian carabinieri, l1 identity solution, analysys iq, University of Konstanz, tno, foi, Ulster University, Guardia Civil, nicc incc, Finnish police, Swedish police, University of Modena and Reggio Emilia, University of Dublin.
Funding: Finnish minister of finance
Principal investigator: Finnish minister of finance
State: Finland
ADM in action: The AI assistant, AuroraAI, aims to help provide efficient and effective public services by identifying life events and suggesting relevant services to citizens. It involves bringing together data collected in various parts of the public sector to proactively develop services for citizens. The system automatically identifies life events (e.g., moving to a new city, retirement, loss of employment) and suggests services accordingly. It acts as a recommender system for public services, aiming for timeliness, personalization, targeting, and automated service provision to increase efficiency.
Partners: Finnish Ministry of Economic Affairs and Employment, NGOS
Link: https://www.ccdriver-h2020.co
Funding: Horizon 2020 European Programme
Total Costs: €4,997,630.00
Principal Investigator: Trilateral Research LTD
ADM in action Development of an online self-assessment metric tool for young people to understand cybercriminal behavior and promote positive pathways. Additionally, the project has created an online self-assessment questionnaire that SMEs, CSOs, and other stakeholders can use anonymously to assess their vulnerability to cybercrime attacks. Users will receive a score and recommendations to address vulnerabilities. For Law Enforcement Agencies (LEAs), CC-DRIVER is developing tools to collect evidence using cloud forensic medicine and to investigate and mitigate cybercrime operations.
Partners: Trilateral Research, withSecure, Foundation for Research and Technology, Software Imaginarium and Vision, Valencia Local Police, Polícia Judiciária, University of Applied Sciences for Public Services in Bavaria, University of Lausanne, Center for Security Studies, University of East London, Information Security Forum, Privanova, Hellenic Police
Link https://www.connexions-project.eu
Founding: Horizon 2020 Europe Programm
Total costs: €4,999,390.00
Principal Investigator: Ethniko kentro erevnas kai technologikis anaptyxis
Country: Greece
ADM in action: The objective is to develop and demonstrate next-generation services for detection, prediction, prevention, and investigation. These services will be based on the multidimensional integration and correlation of heterogeneous and multimodal data, providing relevant information to various stakeholders interactively and tailored to their needs through augmented and virtual reality environments. The main activities include extracting information from multilingual and multimedia content, propagating and analyzing IoT data, multimodal integration and retrieval of information, correlation and distribution of multimodal information, creating engaging environments to enhance situational awareness and investigative capabilities, a real-time operational command center (OpCom), system development and integration, defining user requirements, field demonstrations, user evaluation and training, legal and ethical considerations and social impact, exploitation and sustainability, dissemination and collaboration.
Partners: motorola solutions israel ltd, sheffield hallam university, nuromedia gmbh, universidad pompeu fabra, police service of northern ireland, hochschule fur den offentlichen dienst in bayern, ministério da justiça, engineering – ingegneria informatica spa, serviciul de protectie si paza, bayerisches staatsministerium des innern, kentro meleton asfaleias - centre for security studies, software imagination & vision srl
Link: https://copkit.eu
Funding: Spanish Ministry of Defense
Principal Investigator: Ingeniería de Sistemas para la Defensa de España S.A. S.M.E. M.P.
Country: Spain
ADM in action: The COPKIT project has developed data-driven policing technologies to assist Law Enforcement Agencies in analyzing, investigating, mitigating, and preventing the use of new information and communication technologies by organized crime and terrorist groups. The COPKIT toolkit supports an Early Warning/Early Action methodology to explain crime evolution, identify trends, send alerts about new risks, and aid decision-makers in developing preparedness, mitigation, prevention, and other security policies. The project’s components are categorized into six ecosystem phases: data collection, information extraction, information enrichment, knowledge discovery, assessment, and forecasting. An innovative Human Machine Interface (HMI) supports analysts through visual analytics. Additionally, COPKIT has facilitated the integration of tools into existing IT systems and the adoption of new tools. All tools were developed with a focus on ethical, legal, and societal aspects to ensure they are ethically acceptable and socially desirable.
Partners: Isdefe, Thales, IBM, Trilateral Research, Legend Technologies, Universidad de Granada, Kemea, Law and Internet Foundation, AIT, Policía Nacional
Link: https://project-crest.eu
Funding: Horizon 2020 European Programme
Total costs: 6.999.078,75€
Principal Investigator: Serviciul de protectie si paza
State: Romania
ADM in action: CREST aims to enhance the effectiveness and efficiency of Law Enforcement Agencies' (LEAs) intelligence, operations, and investigation capabilities by automating the detection, identification, assessment, fusion, and correlation of evidence from diverse multimodal data streams. These data sources include the Surface/Deep/Dark Web, social media, IoT devices, surveillance cameras, and seized devices. To achieve this, CREST is developing an innovative platform for prediction, prevention, operation, and investigation, focusing on multidimensional integration and correlation of these data streams and delivering relevant information interactively to various stakeholders.
Partners: Serviciul de Protectie si Paza, Centre for Research and Technology Hellas, Motorola Solutions, Centric, EVERIS, Lorem University of Technology, SIMAVI Romania, Maynooth University, COPTING Gmbh, Cyberlens,National and Kapodistrian University of Athens, ROBOTNIK, Universität Wien Department of Digitalisation and Innovation in Law, IN MANIBUS MEIS, Munich Police Department, Center for Security Studies,Hochschule fur den offentlichen Dienst in Bayern, State Protection and Guard Service of the Republic of Moldova, Polícia Judiciária,Hellenic Police,Politsei- ja Piirivalveamet, Defence Institute “Professor Tsvetan Lazarov”, Victim Support Europe
Link: https://www.h2020-dante.eu
Funding: Horizon 2020 European Programme
Total cost: 6.206.216,25€
Principal Investigator: Engineering
State: italy
ADM in action: The system automatically translates texts from various languages into English and summarizes the information using natural language processing algorithms, which process and extract the most relevant information. It can recognize coded messages and perform detailed linguistic analysis between and within texts. Groups, leaders, and suspicious individuals can be identified, analyzed, and monitored at various levels, including sociological, criminological, and psychological, to identify behavioral patterns. The project develops a series of innovative services for the automated detection and analysis of online multimedia and multilingual content, based on existing solutions and prototypes from project partners or available as open source software. A TRL-7 system prototype is being developed, integrating and leveraging the capabilities offered by these services. The system will provide intelligence and collaboration applications and analysis services for law enforcement agencies. DANTE’s ultimate goal is to equip analysts with tools to combat terrorist activities, thereby contributing to a safer civil society.
Partners: Engineering – Ingegneria Informatica Spa, Expert System Iberia S.L.U., Centre for Research and Technology Hellas / Information, Technologies Institute, Ciaotech S.r.l., RiSSC, PROMT GmbH, Vocapia Research, United Technologies Research Centre, Ireland Ltd., AGNITIO S.L., AIT Austrian Institute of Technology GmbH, Trilateral Research Ltd, Fundación Deusto –, DeustoTech,Pragsis Bidoop, KU Leuven Centre for IT and IP Law – IMEC, GUARDIA CIVIL, Polícia Judiciária, Home Office Centre for Applied Science and Technology (CAST)
Tool Users: Interpol, Europol, Eurocrime, Public security portal, ENFSI
Link: https://www.darleneproject.eu
Funding: Horizon 2020 European Programme
Total cost: 6.954.860,00€
Principal Investigator: Ethniko kentro erevnas kai technologikis anaptyxis
ADM in action: DARLENE is designed as an augmented reality (AR) system for Law Enforcement Agencies (LEAs) and first responders to enhance physical and mental processing capacities, improving situational awareness and supporting decision-making under stress. Situational awareness is crucial for the safety of officers and community members, and DARLENE aims to improve this through advanced AR technology. The system employs the OODA Loop—Observe, Orient, Decide, Act—which emphasizes agility in decision-making to outmaneuver opponents. DARLENE uses real-time AR tools to rapidly process visual information and respond to threats. The technologies will enable LEAs to develop a common operating picture by collecting data from various sources and analyzing it with machine learning, providing rapid, coordinated assessments and revealing blind spots. Combining AR, machine learning, and 5G technology, DARLENE offers innovative tools for European LEAs to gain tactical advantages and stay ahead of terrorists and other adversaries.
Partners: CERTH, ICS-Forth, CTTC, Ku Leuven Centre, KEMEA, Thales, Trilateral Research, eBos, Youbiquo,Basque Government-Security Department, Policia Judiciaria, HfoeD, Policia local de Valencia, Cyprus Police, Lithuanian Police
Funding: Finnish Government
Principal investigator: Fujitsu
State: Finland
ADM in action: The project uses machine learning methods to help identify factors underlying social exclusion of young adults and predict associated risks. The predictive model provides an overview of risk factors for social and healthcare professionals. The model is derived from pseudonymized data from young adults using Eksote’s services and predicts social exclusion outcomes defined by Eksote’s professionals. The legislation allows only the use of non-identifiable, pseudonymized data, so the model produces a general list of risk factors rather than identifying individual young adults at risk.
Partners: Avaintec
Funding: Estonian Government
Principal Investigator: Estonian Health Insurance Fund & World Bank
State: Estonia
ADM in action: Risk-based management model for healthcare aimed at increasing the integration of health services. This model helps family physicians identify patients with multiple chronic illnesses who would benefit from additional preventive care, counseling, and follow-up to improve their quality of life. Neglecting these patients can lead to health deterioration and increased healthcare costs, such as avoidable hospitalizations and duplicate studies. Between 2016 and 2017, EHIF and the World Bank conducted a pilot project in Estonia to refine a clinical algorithm that identifies patients likely to be hospitalized based on their medical conditions. This algorithm uses data from EHIF’s medical invoice database to predict which patients might need more intensive care. The pilot project emphasized the need to prioritize patients based on behavioral data (e.g., prescription adherence) and social characteristics (e.g., social vulnerability), which were not yet fully integrated into the patient selection process.
Link: https://exfiles.eu
Funding: Horizon 2020 European Programme
Total cost. 6.954.86000€
Principal Investigator: TECHNIKON FORSCHUNGS UND PLANUNGSGESELLSCHAFT MBH
State: Austria
ADM in action: To extract information from encrypted devices, a holistic approach involving both software and hardware is necessary. The goal is to access protected evidence by combining semiconductor industry knowledge with software exploitation techniques. EXFILES aims to categorize smartphones used by criminals, enhance existing tools for reverse engineering specific mobile devices, and integrate software and hardware methods to create advanced solutions. The project seeks to make evidence extraction from encrypted smartphones affordable and practical, improve law enforcement capabilities related to encryption, and increase information sharing. It involves stakeholders from various domains and provides guidelines and recommendations for lawmakers and law enforcement agencies, with evaluations based on real use cases.
Partners: TECHNIKON, CEA-Leti,French Gendarmerie Cyberspace Command,Bundeskriminalamt, Centro Nacional de Inteligencia, CSIC, NFI, RHUL, Discurre BV, Texplained, Université de Lille, Cyber Intelligence S.L., National Criminal Investigation service Norway, Synacktiv
Funding: Municipality of Copenhagen
Principal Investigator: Municipality of Gladsaxe
State: Denmark
ADM in action: The model consisted of a points-based analytical system with parameters to estimate risk. These included aspects such as parental mental health (3000 points), unemployment (500 points), missed medical appointment (1000 points), missed dental appointment (300 points), and divorce. Although the overall purpose of the project was commendable, the way it was implemented was heavily criticized and resulted in the model being put on hold. This meant that Gladsaxe and its partner municipalities were not allowed to roll out the system
Link: https://www.grace-fct.eu
Funding: Horizon 2020 European Programme
Total cost: 6.823.512,50€
Principal Investigator: FUNDACION CENTRO DE TECNOLOGIAS DE INTERACCION VISUAL Y COMUNICACIONES VICOMTECH
State: Spain
ADM in action: GRACE, at its core, focuses on three main objectives. Firstly, it aims to address the volume and analyze the content of online Child Sexual Exploitation Material (CSEM) using technological innovations. Secondly, it strives to provide practical operational value to Law Enforcement Agencies (LEAs) in their efforts to investigate online CSEM. Thirdly, GRACE aims to influence strategic and policy-level decisions by harmonizing EU-wide responses to CSE. In terms of approach, GRACE will utilize machine learning techniques for the referral and analysis processes, while tackling the unique technical, ethical, and legal challenges inherent in combating CSE. The project leverages existing resources at Europol and the consortium's nine Member State LEAs. GRACE's strategy includes delivering early, frequent, and flexible results, prioritizing achievable milestones such as deduplication. A distinctive aspect of GRACE is its development and application of Federated Learning, designed to optimize analysis and information sharing in a manner that respects privacy and security. This approach fosters cooperation among LEAs, enhancing their collective capabilities and leveraging their expertise. Ultimately, GRACE aims to provide its outcomes to Europol and Member State LEAs for unrestricted operational use, thereby contributing to their technological independence and efficacy in combating CSE.
Partners: Vicom, AGS, Federal Police, Centric, CERTH, MIRPN, Cybercrime Research Institute, Cyprus Police, DCPJ, Engineering, EUC, Europol, IGPR, INOV, KGP, L3CE, NICC, Policia Judiciaria, Synelixis, Universidad de Leon, WEB-IQ, ZITiS
Link: http://i-lead.eu
Funding: Horizon 2020 European Programme
Total cost: 3.483.716,00€
Principal investigator: Secure societies - Protecting freedom and security of Europe and its citizens
State: Netherlands
ADM in action: The I-LEAD project aims to enhance the effectiveness, efficiency, and capabilities of law enforcement across Europe through a strategic framework that coordinates individual efforts in criminal investigations. Its goal is to create a consolidated and interconnected network that addresses operational and capacity gaps, expresses common requirements, monitors research and innovation, sets priorities for standardization and policy recommendations, and promotes interaction with industry, research, and academia. The project supports capacity development, knowledge exchange, dissemination of results, and interaction with related networks by identifying and integrating past and present technological and innovation initiatives to strengthen the law enforcement network.
Link: https://h2020-infinity.eu/
Founding: Horizon 2020 European Programme
Total cost: 6.866.503,75€
Principal investigator: Airbus defense and space sas
State: France
ADM in action: The main objectives of the INFINITY project are to revolutionize data-driven investigations through the use of artificial intelligence, machine learning, and big data analytics to enhance the effectiveness of investigations, and to employ modern innovations in virtual reality, augmented reality, and visual analytics to improve the intelligence cycle. Both aspects will be driven by end-users and designed to meet the needs of law enforcement. INFINITY will provide an integrated solution aimed at revolutionizing data-driven investigations by bringing together strong representation from national and supranational agencies with a user-oriented design to directly address the fundamental needs of contemporary law enforcement. This will involve synthesizing the latest innovations in virtual and augmented reality, artificial intelligence, machine learning, big data, and visual analytics.
Partners: Airbus, BHFOD, Centric, ITI, CNRS, DFKI, Engineering, Europol, GUCI, Hellenic Police, Kemea, Police Service of Northen Ireland, Police zone Atewep, University of Vienna, Universidad politecnica de Madrid, Vicom, VTT
Tool users: European Police
Link: https://inspectr-project.eu
Founding: Horizon 2020 European Programme
Total cost: 6.997.910,00€
Principal Investigator: University College Dublin & University of Ireland
State: Ireland
ADM in action: The principal objective of INSPECTr is to develop a shared intelligent platform and an innovative process for gathering, analyzing, prioritizing, and presenting key data to support crime prediction, detection, and management across local, national, and international agencies. This data will come from free and commercial digital forensic tools, supplemented by online resource gathering. The platform will facilitate the ingestion and homogenization of both structured and unstructured data with increased automation, allowing for interoperability between different data formats. Using various knowledge discovery techniques, investigators will be able to visualize and bookmark important evidential material and export it to an investigative report. In addition to providing basic and advanced cross-correlation analysis with existing case data, this technique will aim to improve knowledge discovery across exhibit analysis within a case, between separate cases, and ultimately between interjurisdictional investigations. INSPECTr will employ big data analytics, cognitive machine learning, and blockchain to significantly enhance digital and forensic capabilities for pan-European law enforcement agencies. INSPECTr aims to reduce the complexity and costs for law enforcement agencies and related actors in using cutting-edge analytical tools proportionally and in line with relevant legislation. The final developed platform will be freely available to all law enforcement agencies. INSPECTr seeks to increase EU security while adhering to Responsible Research and Innovation principles and national and European research ethics requirements in a manner that respects and preserves civil liberties.
Partners: AGS, UCD-CCI, IGPR, LSP, PSNI, GN, BFP, CNR, ILS, MoJN, RUG, PHS, EPBG, Trilateral Research, Universitè de Lausanne, eBOS, Siren, VLTN
Funding: Belgian Government
Principal Investigator: Ministry of Home Affairs & Ministry of Justice
State: Belgium
ADM in action: In Belgium the force is divided into the Administrative Police and the Judicial Police. While the Judicial Police is involved in investigations and judicial proceedings, the Administrative Police performs various functions, from enforcing road traffic laws to maintaining public order. These functions are defined by the 1992 law on the functioning of the Police. This law established a complex system of databases, where all information – including personal data – collected by the Police must be stored. Depending on whether they are local or federal, administrative or judicial, these databases are accessed by officers and administrative staff through multiple applications. The iPolice project aims to address the fragmentation of these databases and the proliferation of platforms by consolidating them into a single solution. This unified platform will significantly enhance the efficiency of the Belgian Integrated Police by providing quicker access to more and better-organized information. Additionally, the solution includes intelligence-led analysis features that will enable the Police to analyze their data more effectively while also adhering to data protection rules.
Funding: Estonian Government
Principal Investigator: The Public Transport and Traffic Management Department in Tallinn
State: Estonia
ADM in action: This solution aims to monitor traffic flow in Tallinn, particularly the daily influx and outflow of cars. The data gathered helps address issues such as parking problems and road construction planning. Initially, three cameras were installed at intersections in Tallinn to test the system and train the software to count vehicles passing by. The machine vision algorithm categorizes and counts buses, cars, trucks, and motorcycles. Future enhancements aim to extend this capability to pedestrian counting. However, the accuracy of the results can be influenced by adverse weather conditions, such as fog, and obstructions like dirt on the cameras.
Link: https://www.lasie-project.eu
Founding: European Union
Total cost: 11.352.974€
Principal investigator: Engineering
State: Italy
ADM in action: The LASIE project aims to design and implement an open and expandable framework that supports analysts in managing and analyzing large amounts of heterogeneous forensic data. LASIE will significantly increase the efficiency of current investigative practices by providing automated analyses of forensic data acquired from a variety of sources, including CCTV surveillance content, confiscated desktops and hard drives, mobile devices, the Internet, social networks, handwritten and calligraphic documents.
Link: https://letscrowd.eu/
Funding: European Union
Total cost: 2.919.307,50€
Principal Investigator: Etra investigarion y desarrollo
State: Spain
ADM in action: LETSCROWD will overcome challenges preventing the effective implementation of the European Security Model with regards to mass gatherings. This will be achieved by providing the following to security policy practitioners and, in particular, Law Enforcement Agencies (LEAs): a dynamic risk assessment methodology for the protection of crowds during mass gatherings centered on human factors to effectively produce policies and deploy adequate solutions; a policy-making toolkit for long-term and strategic decision making of security policy makers, including a database of empirical data, statistics, and an analytical tool for security policy modeling; and a set of human-centered tools for LEAs, including real-time crowd behavior forecasting, innovative communication procedures, semantic intelligence applied to social networks and the internet, and novel computer vision techniques. LETSCROWD’s impact will be measured through practical demonstrations involving seven LEAs and relevant emergency services units.
Partners: ETRA I+D, Policia Municipal de Madrid, University of applied sciences police affairs, Crowd Dynamics International, Deep Blue, European Emergency number association, Expert System spa, Local Police Voorkempen, Ministry if the interior polizia di stato dipartimento di sicurezza pubblica, Zenabyte srl, Railsec ltd, Ministry of internal affairs ministrerul afacerilor interne, Universidad de Cantabria, Università di Cagliari, Ertzaintza, Pluribus One
Link: https://lion-dc.eu
Funding: European Union
Total cost: 836.860,38€
Principal investigator: Mykolo Romerio Universitetas
State: Lithuania
ADM in action: Lion DC is a cross-border exercise based on real cases of drug-related crime, involving law enforcement professionals and enabling knowledge exchange with academia. Objective 1 is to create a functional online drug investigation community, comprising professionals and academic participants, by extending the SENTER, ENTER, and i-LEAD initiatives. Objective 2 is to develop new investigative methods using available OSINT, Dark Web, and Cryptocurrencies Investigation technologies. This will provide an opportunity to share analytical expertise, methodologies, and tools, as well as to evaluate and classify available tools based on their feasibility and utility.
Partners: Mykolo Romerio Universitetas, DITSS, ICT Academy, KEMEA, L£CE, Lietuvos Muitine, Ministero della difesa Olandese, TNO
Link: http://www.magneto-h2020.eu
Funding: Horizon 2020 European Programme
Total cost: 5320.475,00€
Principal Investigator: Erevnitiko Panepistimiako insititouto systimaton epikoinonion kai ypologiston
State: Greece
ADM in action: MAGNETO will test and demonstrate its developments on five representative and complementary use cases (types of crime) under real operational conditions in the facilities of eleven different Law Enforcement Agencies (LEAs). This will be done by keeping them continuously in the production cycle, adopting an agile implementation methodology, and a multidisciplinary scientific approach, combining researchers with exceptional backgrounds, officials with high-level operational know-how in law enforcement, experts recognized for legal and ethical compliance with EU and national standards, and qualified training experts for the development of innovative programs. The main objectives of the project include: Increasing LEAs’ capabilities to manage and analyze large amounts of heterogeneous data to analyze intelligence, extract evidence, and support police operations; Developing technical measures to effectively anticipate both the growth in volume and types of police data and to accurately monitor crime developments; Designing and composing an open framework and developing powerful tools to provide greater innovation capacity in Europe and more efficient information exchange between operators and other users. MAGNETO’s technologies and solutions will enable LEAs to efficiently process large volumes of heterogeneous data and gather reliable, court-proof evidence. Legal and ethical compliance is a cornerstone of the project. In collaboration with technical partners and end-users, human rights law, criminal procedure, and data protection principles are translated into concrete design strategies to implement safeguards directly into the system architecture.
Funding: Estonian Government
Principal Investigator: Estonia’s Unemployment Insurance Fund
State: Estonia
ADM in action: When a citizen registers as unemployed on the EUIF website, their data is checked, and if accurate, they are registered as unemployed. The system uses AI to verify an applicant’s data across various databases and then determines which document to send to the applicant. The EUIF also employs Automated Decision Making (ADM) to determine eligibility for unemployment aid or unemployment insurance aid, including the amount and duration. According to Erik Aas, an EUIF council member, 50% of these decisions are made entirely through ADM. Additionally, the EUIF plans to start predicting how long specific individuals will remain unemployed, aiming to leverage all the data it possesses. The EUIF development plan for 2019-2022 aims to increase the impact, quality, and accessibility of services by using a profiling model. This model helps identify the risk of long-term unemployment for each client and determines the appropriate counseling channel, frequency, and specific services.
Link: http://www.prevision-h2020.eu/
Funding: Horizon 2020 European Programme
Total cost: 8.001.180,00€
Principal Investigator: Erevnitiko Panepistimiako insititouto systimaton epikoinonion kai ypologiston
State: Greece
ADM in action: PREVISION partners will leverage their capabilities, expertise, and prior research, alongside established and evolving European standards and best practices. This approach allows them to concentrate resources on the project’s innovative aspects. The execution strategy revolves around an iterative development methodology, ensuring frequent software releases for testing and evaluation by LEA and practitioner end-users. This process aims to keep them actively involved in the production cycle. The PREVISION Platform will undergo deployment in 10 distinct demonstrations overseen by consortium LEAs and practitioners.
Partners: ICCS, Fraunhofer IOSB, etra I+D, ITTI sp.zo.o, IfmPt, BPTI, CERTH, Politecnica University of Valencia, SIMAVI, CNRS, Space hellas, CUING, MCA, PARCS, Trilateral Research, Maastricht University, Catalink Limited, System Factory, ENSP, Departamento de Seguridad Gobierno Vasco, FNTT, Hellenic Police, HfoeD, KEMEA, State Protection and Guard Service, PPM, PSNI, SSP, HisoMA
Link: https://www.projectproton.eu/
Funding: European Union
Total cost: 4.094.506,68€
Principal Investigator: Università Cattolica del Sacro Cuore
State: Italy
ADM in action:Advancement of understanding of the social, psychological, and economic factors leading to organized crime and terrorist networks, including their connection to cybercrime and cyberspace. Development of initial ABM simulations on the processes leading to organized crime and terrorist networks, including recruitment into these networks. These simulations have enabled evidence-based policies and will foster further innovation. Development of the first policy support tool (PROTON Wizard) providing easy access to cutting-edge scientific research for non-scientists, supporting improved policies at international, national, and local levels. PROTON-S, a suite of distinct ABM models simulating processes leading to recruitment in organized crime and terrorist networks, and testing the impact of specific policy interventions. PROTON Wizard, a user-friendly web interface allowing policymakers and stakeholders to navigate PROTON-S results and test various scenarios easily. The goal is to enable policymakers to access simulation outputs through a clear interface without specialized assistance. Policy recommendations for recruitment into organized crime and terrorist networks.
Partners: Transcrime
Link: https://ramses2020.eu
Funding: Horizon 2020 European Programme
Total cost: 3.532.000,00€
Principal Investigator: Politecnico di Milano
State: Italy
ADM in action: The system will extract, analyze, link, and interpret internet-derived information on financially motivated malware. This project will rely on disruptive Big Data technologies to first extract and store, and subsequently search for patterns of fraudulent behavior within vast amounts of structured and unstructured data. It aims to develop effective guidelines and collaborative methodologies for law enforcement investigations. Development of a series of tools for Internet Forensics. Demonstration of the impact of the RAMSES platform through various pilot exercises in different countries, training campaigns, and awareness initiatives. Enhancement of forensic analysis tools applied to internet-connected devices in Europe.
Partners:nPolitecnico di Milano, Policia Judiciaria, University of Kent, RISSC, Universidad Complutense de Madrid, College of the Bavarian Police, Trilateral Research, Tree Technology, Belgian Federal Police, CISPA, Cuerpo Nacional de Policia
Tool users: Belgian Police, Spain Police, Portuguese Police, German Police
Principal Investigator: University of Helsinki
State: Finland
ADM in action: The machine learning model is built using data from past human-made content removal decisions. The system can be adjusted over time, such as by altering the model’s sensitivity to posted messages. Suomi24 automatically sends moderation requests to Utopia, and their AI model either accepts or rejects the content. In borderline or new cases, the model may request a human moderator to review the post. The model is kept up-to-date using new samples generated by human moderators.The automated moderation process is a combination of Utopia’s tool, Suomi24’s technical systems, and human work.
Partners: Suomi24, Utopia Analytics
Link: https://roborder.eu
Funding: European Union
Total cost: 8.922.410,03€
Principal Investigator: Ethniko kentro erevnas kai technologikis anaptyxis
State: Greece
ADM in action: ROBORDER aims to develop and demonstrate a comprehensive autonomous border surveillance system using unmanned mobile robots, including aerial, water surface, underwater, and ground vehicles (UAVs, USVs, UUVs, and UGVs). These robots will incorporate multimodal sensors within an interoperable network. The system's main objective is to enhance detection capabilities for early identification of criminal activities and environmental incidents in border and coastal areas, including marine pollution events. The robots will operate both individually and in swarms, equipped with adaptable sensing technologies suitable for various operational environments. The sensor network will include static border surveillance radars and mobile sensors such as passive radars, RF-signal sensors, thermal and optical cameras. These sensors will provide detailed situational awareness, allowing efficient command and control operations. Supplementary technologies will ensure robust communication links between command units and the robots. Advanced command and control functionalities will enable operators to integrate and analyze heterogeneous sensor data swiftly, supporting informed decision-making and remote operational actions tailored to specific situations.
Partners: CERTH, FHR, EASS, VTT, EVERIS, PSNI, GNR, NATO Science & Technology Organization, HNP, ROB, SPP, Elettronica GmbH, Hellenic Ministry of Defense, CENTRIC, AdSP-MTS, OceanScan, BDI,Copting GmbH, University of Athens, CSEM, CNIT, Polícia Judiciária, Cyberlense LTD, Romanian Border Police, Everis, Tekever
Link: https://www.roxanne-euproject.org
Funding: European Union
Total cost: 6.999.458,75€
Principal Investigator: Fondation de l'institut de recherche IDIAP
State: Switzerland
ADM in action: The ROXANNE project aims to develop an advanced tool for law enforcement agencies that combines speaker data mining with link analysis to efficiently track criminals and terrorists. Leveraging the conversational nature of speech data, the project will focus on reliable diarization to determine speakers during conversations, even when acoustic evidence is unreliable. It will also integrate call content, using commercial speech recognition engines to rapidly convert speech to text and identify details such as names and locations. Metadata analysis will be crucial, aiming to extract information such as age, gender, and the environment of calls to better understand criminal networks. The project will also address challenges like data anonymization, using innovative techniques to overcome these limitations. The ultimate outcome will be a prototype system capable of handling large volumes of voice data and metadata, analyzing them in an unsupervised manner, and presenting results in formats compatible with standard investigative software solutions. ROXANNE aims to significantly enhance law enforcement's investigative capabilities by integrating advanced technologies to identify persons of interest and simplify criminal network analysis, with the goal of deploying the system in real-world cases and ensuring compliance with EU and INTERPOL legal and ethical frameworks.
Partners: Idiap, Phonexia s.r.o, BUT, Hensoldt, University of Hannover, University of Saarland, Università Cattolica del Sacro Cuore, Transcribe, Trilateral Research, AEGIS, Airbus, ADITESS, ITML, Interpol, Lithuanian Forensic Institute, Romania Minister of Interior, Police Service of North Ireland, Police d Czech Republic, Croatia Police, KEMEA, Israel National Police, NFI, An Garda Siochana, Hellenic Police, ZITIS
Funding: Estonia government
Principal Investigator: Estonia municipal governments
State: Estonia
ADM in action: This tool helps case managers (such as social workers, child protection officers, and youth workers) identify young people aged 16-26 who are not in education, employment, or training (NEET). It is designed to support these individuals in returning to education or entering the labor market. Case managers or municipal government employees can access information about young people living in their specific locality through the YGSS. However, not all municipalities have joined the program. To participate, municipal governments must submit an application to the Municipal Government Information System for Social Services and Benefits (STAR) (Noortegarantii | Sotsiaalministeerium o. J.). Once enrolled, case managers or municipal employees can view a list in the STAR system of all NEET youth registered as residents in their area who, according to other registry data, need assistance. This information is sourced from nine different registries, creating a comprehensive dataset that includes details such as the young person’s address, email, phone number, educational background, and language of study (Estonian or Russian). Young people have the right to opt out of having their data analyzed by a case manager, and consent is required for any data processing. The case manager then contacts the identified young people, sending a letter or SMS asking them to complete a survey. The Youth Guarantee Support System request is automatically triggered twice a year, on March 15th and October 15th.
Link: https://shotpros.eu
Funding: Horizon 2020 European Programme
Total cost: 5.059.843,75€
Principal Investigator: Usecom the usability consultants
State: Austria
ADM in action: The SHOTPROS project addresses the evolving challenges faced by European street patrol police officers, who increasingly find themselves as first responders in high-risk and stressful situations. These situations are growing in frequency and complexity, requiring enhanced decision-making and action (DMA) performance. SHOTPROS aims to investigate how psychological and contextual human factors influence police officers' DMA behavior under stress. It will develop a Virtual Reality (VR) solution to experimentally assess these factors and design a training curriculum rooted in human factors (HF) principles. This HF-based training, combined with VR simulations, aims to improve DMA-SR performance, leading to more effective decision-making, reduced use of force, and minimized casualties and collateral damage in critical incidents.
Partners: Berlin Police, NCCN, Nacional Police of the Netherlands, LAFP NRW, Romanian Ministry of Internal Affair, Swedish Police, Austrian Institute of Technology, Katholieke Universities Leuven, Re-liOn Group, Ruprecht-Karls-Universitat Heidelberg, Vrije Universiteit Amsterdam, VESTA
Principal Investigator: Skillific
State: Estonia
ADM in action: Recruitment service using machine learning to help employers find potential employees by analyzing the required skills and knowledge. The service uses the European Skills, Competences, Qualifications, and Occupations (ESCO) classification system to match candidates with job profiles. Skillific’s AI application searches through web data, partner data, and public databases to find and profile potential candidates, assessing their suitability for various roles. Currently, the software is still in the testing phase due to insufficient high-quality training data. The Skillific database contains about 400,000 user profiles and a smaller number of job profiles. The machine learning algorithm predicts candidate suitability and improves data storage and collection efficiency, making the recruitment process less resource-intensive and more effective.
Funding: Estonian Government
Principal Investigator: Sotsiaalkindlustusamet
State: Estonia
ADM in action: This initiative leverages the state's existing information about each citizen from birth, allowing subsequent services, such as family benefits, to be activated automatically. It represents the first event-based service in Estonia, with the SKA aiming to transition fully to application-free services.Once a birth is registered in the population register and the child is named, the Social Security Agency sends an email to the parents. Upon confirming receipt of this notification, parents can receive family benefits.
Partners: Nortal
Link: https://www.spirit-tools.com/index.php
Funding: Horizon 2020 European Programme
Total cost: 4.998.656,25€
Principal Investigator: Lutech s.p.a.
State: Italy
ADM in action: The SPIRIT project introduces a novel approach to developing, testing, training, and evaluating a scalable prototype system for privacy-preserving intelligence analysis in identity resolution. Initially, SPIRIT will focus on specific use cases provided by End User Partners, implementing a privacy by design framework to create an early prototype installed at each partner's office. The system will utilize anonymized datasets and advanced functionalities to securely analyze open source data within each partner's firewall and security systems. SPIRIT aims to deliver a suite of tools for acquiring, analyzing, modeling, and visualizing multimodal, multilingual, and multimedia data (including text, audio, video, and images) from diverse sources such as surface web, deep web/dark nets, and social networks. The platform will enable the creation of social graphs depicting relationships between named entities of varying types, facilitating social and criminal network analysis. Throughout development, SPIRIT emphasizes privacy by design and privacy through design principles to assist law enforcement agencies (LEAs) in navigating the complex landscape of identity resolution ethically and transparently. The project will also develop a comprehensive training course accredited by universities for LEAs and personnel involved in this domain. Additionally, a distance-learning course will be created for developers and researchers, emphasizing the ethical and technical responsibilities associated with creating and utilizing such tools.
Partners: Nydor System Technologies, A E Solutions, Live Srl, Singular Logic, London Metropolitan University, ROCU, STAD Antwerp, Police and Crime Commissioner for Thames Valley, Innova Integra Limited, MInistarstvo Unutrasnjih Poslava Republike Srbije, Universitat Autonoma de Barcelona, Hellenic Police, Linkopig University, European Center of Psychology Investigation Criminology, Fraunhofer IDMT
Link: https://www.starlight-h2020.eu
Funding: Horizon 2020 European Programme
Total cost: 18.932.200,63€
Principal Investigator:Commissariat a l' energie atomique et aux energies alternatives
State: France
ADM in action: STARLIGHT is committed to achieving its strategic goals through a human-centric approach to AI development. This involves carefully addressing ethical and legal implications to ensure that AI tools are responsible, meeting societal needs while mitigating negative impacts and embodying ethical values. Decision-making processes will be transparent and understandable, supporting the admissibility of resulting evidence in legal contexts. The project aims for technological robustness and compliance with all applicable laws and ethical principles, ensuring accountability through assessable algorithms and design processes aligned with existing EU frameworks and regulations. To expedite its objectives, TARLIGHT builds upon insights and practices from previous and concurrent projects. Leveraging established systems, best practices, and evolving AI components, the project integrates ethical and legal frameworks to guide its development. Detailed use cases serve as benchmarks for implementing effective solutions. The project employs co-design and co-creation methodologies, engaging LEAs and technology providers in iterative research, development, and testing cycles. This collaborative approach identifies gaps, requirements, and challenges, fostering flexibility and adaptability amidst evolving AI technologies and LEA priorities. An open-source approach is central to TARLIGHT's strategy, promoting broad adoption, adaptation, and scalability of its technologies and solutions across national and EU levels. Regarding privacy and data protection, TARLIGHT adheres to a "privacy and data protection by design and by default" principle. This approach ensures that AI tools, services, and the STARLIGHT framework are developed in compliance with GDPR, the Law Enforcement Directive (LED), and forthcoming AI regulations. Continuous integration of privacy measures throughout the project lifecycle upholds data security and legal standards.
Partners: CEA, AMS, AIT, BMI, BPOL, IRIT, CFLW, DFKI, ENG, Ethical and Legal Plus, CERTH, Europol, VICOM, Hellenic Police, Herta security, IANUS, INOV, ITTI, KU Leuven, KEMEA, Fondazione Links, French MInistry of Interior, Polícia Judiciária, Spanish Ministry of Interior, Italian Ministry of Interior, Finnish Police, NICC, NFI, Pluribus One, Belgian Federal Police, Swedish Police, Czech Republic Ministry of the Interior, Swedish Police, Estonian Police, CENTRIC, VTT, Thales, National Police of Netherlands, Tilde, Swedish Defense research Agency, Voice Interaction, Polytechnic University of Madrid, Web IQ, XXII, ZITIS, Belgian Police
Principal Investigator: Keila Consumer Co-operative
State: Estonia
ADM in action: Self-service check-in desk equipped with age-verification technology to facilitate the purchase of age-restricted products. This system utilizes Strongpoint self-checkout software and the Yot Digital Identification Platform. When purchasing age-restricted items, customers undergo an age verification process at the self-checkout desk before proceeding to payment. For tobacco products, the customer selects the item at the self-service checkout and undergoes a preliminary face detection process, which issues a voucher allowing access to a tobacco vending machine. The vending machine then performs another age check before dispensing the product. No additional documents are required to verify the customer's age. This solution is believed to be the first of its kind in Europe.