1. Introduction
Antisemitic hate crimes constitute a persistent and evolving threat across European societies. Their challenge for criminal-justice systems lies not only in their legal classification but in the operational requirement to recognise, respond to, investigate and adjudicate them coherently across both physical and digital environments. Antisemitism today rarely manifests as an isolated offline incident; it is increasingly entangled with online ecosystems of hate speech, coordinated narratives, and symbol-based dehumanisation.1,2,3 Offline events (protests, anniversaries, geopolitical crises) are routinely followed by spikes in online antisemitic activity, and online narratives in turn prefigure offline victimisation.4,5
This dual-environment character poses three intertwined operational problems. First, recognition requires more than legal definitions; it depends on contextual and symbolic literacy that allows first responders, investigators and prosecutors to distinguish bias indicators from ambient noise. Second, response requires coherent action across a chain of actors (police, prosecutors, judicial authorities, community organisations, online platforms), whose procedures, mandates and time-frames are not naturally aligned. Third, investigation and adjudication must produce evidentially robust outcomes without inflicting secondary victimisation on those already harmed.6,7
The EXODUS project (CERV programme, Grant Agreement No. 101144655), coordinated by the Hellenic Police with the Center for Security Studies (KEMEA) leading several activities, was conceived to address precisely this complex of problems within the Greek legal-operational context and in alignment with established European reference frameworks (notably OSCE/ODIHR and the EU Digital Services regime). The project produced a set of analytical outputs of which several (the conceptual and contextual framework,8 the training needs analysis9 and the handbook on antisemitic symbols and recognition)10 are publicly available; others contain operational practice and remain consortium-restricted. This article asks: how can a coherent 3i-based framework for responding to antisemitic hate crimes across physical and digital environments be operationalised, and how can such a framework be grounded in the EXODUS project’s public analytical work in dialogue with international standards and the Greek legal-procedural order?
Three sub-questions structure the analysis: (i) What conceptual and contextual foundations underpin the framework’s recognition component? (ii) What institutional roles, operational needs and coordination challenges must the framework address, and how do they map onto the Greek legal-procedural framework? (iii) What competency and training gaps, and what underlying adult-learning principles, explain the framework’s emphasis on standardisation, common language and practitioner guidance?
The article proceeds as follows. Section 2 sets out the methodology of the synthesis. Section 3 reconstructs the conceptual, operational and training gaps that motivate the framework. Section 4 presents the proposed 3i-based operational framework, clarifies the terminology adopted (formally Identification–Intervention–Investigation, drawing on but distinct from Ratcliffe’s Interpret–Influence–Impact) and locates the framework within the Greek legal-procedural context. Sections 5–7 analyse each pillar in turn; Section 7 also addresses the adult-learning grounding of the framework’s training rationale and its relation to the OSCE/ODIHR TAHCLE tradition. Section 8 examines the cross-cutting victim-sensitive dimension. Section 9 discusses transferability and limitations and Section 10 concludes.
2. Methodology
This article is a synthetic essay rather than an empirical study. It builds on the principles of qualitative framework analysis11 (iterative reading, conceptual mapping and structured synthesis across heterogeneous sources) applied to the operational problem of antisemitic hate-crime response. The analytical strategy is reconstructive and proposive: it makes visible how a coherent 3i logic can be operationalised for this problem, drawing on multiple bodies of material rather than on a single primary text.
Four bodies of material inform the synthesis:
• the intelligence-led policing (ILP) tradition exemplified by Ratcliffe12 and codified in the OSCE TNTD/SPMU Intelligence-Led Policing guidebook;13
• the Greek legal-procedural order, anchored in Article 81A of the Penal Code14 and the institutional architecture for hate-crime response across the Hellenic Police, the Cyber Crime Directorate (also within the Hellenic Police), prosecutorial and judicial authorities, the National Contact Point on Hate Crimes, and the National Commission for Human Rights;
• the publicly available analytical work of the EXODUS project (the conceptual and contextual framework,8 the training needs analysis9 and the handbook on antisemitic symbols and recognition)10 used as developmental background documents that inform the framework’s content;
• the broader scholarly and policy literature on hate-crime criminology, intelligence-led policing, online antisemitism detection, and adult-learning theory.
Materials were read iteratively, with the OSCE/ODIHR corpus on hate-crime and intelligence-led policing serving as the international reference and Article 81A serving as the Greek legal-institutional anchor. The article does not draw on EXODUS deliverables that remain consortium-restricted; analytical claims that the project’s restricted work also addresses are here grounded independently in the public sources cited.
The article does not present new empirical data and does not involve human subjects. No ethical approval was required. Reproducibility is supported by the open availability of all cited sources.
3. Why a framework is needed: conceptual, operational and training gaps
The argument for consolidating the response to antisemitic hate crimes in a coherent framework rests on three diagnosed gaps.
3.1 The conceptual gap: antisemitism as a moving target
The EXODUS conceptual and contextual framework8 maps the conceptual landscape of antisemitism in physical and digital environments. Two findings are central to the framework’s design. First, antisemitism is intrinsically multi-form: religious, ethnic, political, conspiracy-based and Israel-related variants coexist and interact.15 Sharansky’s16 “3D” test (Demonization, Delegitimization, Double Standards) has been incorporated as a tool to distinguish legitimate political criticism of Israel from antisemitism, and is referenced in OSCE/ODIHR practical guidance.17 Second, the working definition of the International Holocaust Remembrance Alliance,18 while widely adopted as a non-binding reference, does not by itself supply the contextual and symbolic literacy practitioners need to recognise antisemitism in real incidents, particularly online, where conspiracy narratives, coded language and visual symbolism evolve rapidly.1,19,20 A coherent framework therefore needs to bridge a definitional anchor (IHRA) with operational indicators (symbols, rhetoric, narrative patterns) and contextual interpretation.
3.2 The operational gap: fragmented response chains
Responsibility for antisemitic hate crimes is distributed across actors with different mandates and time-frames. In Greece, this includes first responders within the Hellenic Police, the Racist Violence Departments and Offices, the Cyber Crime Directorate (also within the Hellenic Police), prosecutors, examining magistrates, courts, community liaison officers, civil-society watchdogs, online platforms and trusted flaggers. Beyond the distribution of roles, the procedural linkage between physical incidents and their online correlates (or vice versa) is rarely codified, producing systematic friction at the interfaces.21 The broader literature documents under-reporting and procedural fragmentation as recurring challenges in hate-crime response across European jurisdictions.7,22
3.3 The training gap: knowledge that does not travel
The EXODUS training needs analysis9 identifies competency and skills gaps that impede effective response across the chain. Three deficits stand out: limited familiarity with antisemitic symbolism and rhetoric beyond the most overt cases; insufficient training in the handling of digital evidence linked to hate content; and uneven awareness of victim-centred procedures.23,7 What knowledge exists tends to remain siloed within specialised units or individual experts rather than being available as a shared vocabulary across the response chain. This finding motivates a framework emphasis on common language, standardised tactics-techniques-procedures (TTPs) and practitioner-oriented tooling, an emphasis with substantive adult-learning underpinnings discussed in Section 7.3.
Taken together, the conceptual, operational and training analyses point not to a missing manual but to a missing connective tissue: a framework able to hold definitional, procedural and competency dimensions in a single coherent structure. This is the function the proposed 3i operationalisation performs.
4. The proposed 3i-based operational framework
4.1 Terminological clarification: from Ratcliffe to operational 3i
A 3i framework for antisemitic hate-crime response requires explicit terminological clarification. The original 3i model formulated by Ratcliffe12 within the intelligence-led policing tradition uses the triad Interpret–Influence–Impact: analysts interpret the criminal environment, intelligence products influence decision-makers, and decisions produce measurable impact on the conditions that enable crime. In operational practice, and most visibly in the OSCE/ODIHR guidance for bias crimes,24,17 this analytical triad has been reformulated as Identification–Intervention–Investigation, a re-articulation that preserves the cyclical, intelligence-led logic of the original while mapping it onto the procedural sequence familiar to law enforcement and judicial authorities. The framework’s underlying ILP logic is itself codified in the OSCE TNTD/SPMU Intelligence-Led Policing guidebook.13
This article therefore deploys 3i in its operational sense (Identification–Intervention–Investigation) while retaining the ILP rationale: intelligence is not collected for its own sake but to enable informed decision-making, dynamic interaction across the response chain, and measurable change in the conditions that allow antisemitic hate crimes to occur. The framework explicitly resists a linear-sequence reading; it conceives 3i as a continuous cycle of understanding, response and feedback.12
4.2 Three framework properties
The proposed framework is structured around the 3i logic but it is not a checklist-style operational manual: its added value lies in three framework properties:
• Continuity: it treats Identification, Intervention and Investigation as phases of a single response cycle, with explicit handover points (e.g., from first responder to investigator, from investigator to prosecutor), rather than as discrete activities.
• Cross-environment scope: each phase is addressed both in its physical and its digital manifestation, with attention to the procedural interfaces that link them (e.g., online evidence collection for an offline incident).
• Whole-chain cooperation: the framework explicitly maps the roles and interactions of police, prosecutors, judicial authorities, the Jewish community, civil-society organisations, online platforms and trusted flaggers, treating coordination as a design requirement rather than an afterthought.
These properties provide the connective scaffolding that allows operational tools to function as a coherent system rather than a collection of detached procedures.
4.3 Legal-institutional anchoring in the Greek context
The framework is anchored in the Greek legal-procedural order. Three reference points are particularly important. First, Article 81A of the Greek Penal Code14 establishes bias motivation, including antisemitic motivation, as an aggravating circumstance, with direct consequences for prosecutorial decisions, charging strategy and sentencing. Recognition of the antisemitic motive is therefore not analytically optional; it is legally constitutive of the offence’s characterisation. Second, the institutional architecture of response is distributed across specialised units: the Cyber Crime Directorate, the Racist Violence Departments and Offices within the Hellenic Police, the National Contact Point on Hate Crimes, and the National Commission for Human Rights (NCHR), each with distinct mandates that the framework must harmonise. Third, transnational cooperation is mediated through the Hellenic Police’s relationships with Europol (notably the EU Internet Referral Unit, EU IRU), Interpol (via iARMS and I-24/7), and the OSCE/ODIHR standards and contact-point network. The framework’s design choices, in particular its emphasis on bias-indicator documentation and on physical–digital interface protocols, are intelligible against this institutional backdrop.
5. Identification: conceptual and contextual literacy
The Identification pillar draws on the public conceptual work8 and the symbol-based handbook.10 Its analytical contribution is to recast identification not as a legal-classification exercise (“does this fit the statute?”) but as a layered recognition task that combines definitional anchors, contextual reading, and symbolic literacy.
5.1 Definitional anchors and their limits
The IHRA working definition18 functions as the framework’s definitional anchor and as a common reference across Member States. The framework treats the IHRA definition as necessary rather than sufficient. Recognition in concrete cases requires translating broad definitional categories into linguistically and visually testable indicators.1,19 The IHRA anchor is therefore complemented with operational indicators drawn from the EXODUS public symbol catalogue,10 from the rhetorical-pattern work in the public conceptual framework,8 from OSCE/ODIHR guidance on distinguishing hate crime from criminalised hate speech,17 and from the OSCE/ODIHR Practical Guide on Understanding Antisemitic Hate Crimes,25 which supplies sample questions and recognition prompts oriented to police practice.
5.2 Symbolic and contextual literacy
The recognition challenge is especially acute online, where antisemitic content blends conspiracy theory, dehumanising rhetoric, visual symbolism and coordinated narratives across platforms.26,27,28 On platforms such as TikTok, repetitive antisemitic motifs are disseminated through multimedia content under coordinated narratives, requiring analysts to read patterns across posts rather than evaluate items individually. The publicly accessible online catalogue of antisemitic and Jewish symbols developed within EXODUS10 allows first responders, civil-society actors and trusted flaggers to consult an authoritative reference when uncertain.
5.3 The role of reporting channels
Identification, in the framework’s logic, includes the reporting infrastructure through which incidents become visible. Official law-enforcement reporting must be situated alongside complementary channels (community organisations, trusted flaggers, platform reporting) and the persistent under-reporting that distorts hate-crime statistics across Europe must be addressed.22 This is significant because the intelligence-led logic of 3i depends on a reasonably representative inflow of identified incidents; without it, the analytical cycle is starved of signal.
6. Intervention: operational coherence and whole-chain cooperation
The Intervention pillar translates identification into action. The discussion here is necessarily at the level of principles; specific tactical procedures used by Greek law-enforcement units remain, appropriately, outside the public domain.
6.1 First response and evidence preservation
First-response procedures in the framework prioritise three simultaneous tasks: protecting the victim, preserving the scene and its evidentiary content, and securing witnesses. The framework requires that hate-crime first response differ from generic incident response in its obligation to record bias indicators contemporaneously and to inform the victim about the subsequent procedural path. Failure at this stage has well-documented downstream consequences: downgrading of the legal characterisation, loss of bias-specific evidence, and erosion of victim trust.29,7,6
6.2 The physical–digital interface
A distinguishing contribution of the proposed framework is its explicit codification of the physical–digital interface within the intervention phase. Offline incidents are increasingly accompanied by online traces (livestreams, perpetrator posts, coordinated commentary, threats directed at the victim) that constitute both evidence and potential continuing harm. The framework requires that the online component be mapped as part of the intervention scope, in coordination with platform reporting channels and, where appropriate, cross-border judicial cooperation mechanisms.21 Open-source intelligence (OSINT) techniques and digital-forensics tooling are integrated as standard methodological resources rather than specialised add-ons.
6.3 EU regulatory anchoring: DSA, Code of Conduct and trusted flaggers
The physical–digital interface relies on a regulatory architecture that has matured rapidly in recent years. The Digital Services Regulation (Regulation (EU) 2022/2065 (DSA)30) imposes due-diligence obligations on online platforms, formalises the trusted-flagger mechanism, and creates structured channels for the rapid notification and removal of illegal content. The revised EU Code of Conduct on countering illegal hate speech online,31 now integrated into the DSA framework, commits participating platforms to standards of speed and quality in the assessment of complaints. The proposed 3i framework treats law-enforcement and civil-society reporting practices as situated within this regulatory environment, with trusted flaggers acting as legitimate intermediaries in the identification–intervention pipeline.
6.4 Whole-chain coordination
The framework’s most ambitious operational contribution is its explicit articulation of “whole-chain” cooperation. Each phase identifies which actor leads, who else is involved, and what is exchanged at each interface.
Table 1 summarises this mapping, reconstructed by the authors on the basis of the Greek institutional architecture and aligned with the OSCE intelligence-led policing approach.
Table 1. Whole-chain cooperation in the proposed 3i framework (actors, lead role and information exchanged across the response cycle).
| Phase | Lead actor(s) | Supporting actors | Information exchanged/outputs |
|---|
| Identification (online) | Trusted flaggers; civil-society organisations; Jewish community references | Platform reporting teams; OSINT analysts; Racist Violence Departments/Offices | Flagged content with contextual annotation (symbols, rhetoric, narrative patterns); platform takedown notices; OSINT briefing notes |
| Identification (offline) | First responders (uniformed police) | Specialised investigators; community liaison officers | Incident report with bias indicators contemporaneously recorded; preservation log; witness contact list |
| Intervention | Specialised investigators; Cyber Crime Directorate (digital components) | First responders; victim-support services; ISPs; EU IRU/Europol where cross-border
| Investigation file with digital evidence integrated; chain-of-custody documentation; victim communication record |
| Investigation (criminal) | Prosecutorial and examining-magistrate authorities | Specialised investigators; expert witnesses; Interpol channels for transnational elements | Indictment file documenting bias motive under Article 81A; admissible digital and forensic evidence; legal characterisation memorandum |
| Adjudication & feedback | Courts (judges) | Prosecutors; defence; victim’s legal representation | Reasoned judgment; sentencing reflecting aggravation; case data fed back into intelligence cycle and training-needs assessment |
The table makes visible what the framework requires of each actor and what flows between them. Three observations are analytically relevant. First, the same incident is iteratively re-described across phases (from “flagged content” through “investigation file” to “indictment” to “judgment”), with the framework’s coherence depending on whether these descriptions remain mutually intelligible. Second, the framework draws on the OSCE intelligence-led policing approach,24,13 in which evidence and assessments flow continuously between analysts, first responders, investigators and prosecutorial authorities. Third, the Jewish community and civil-society organisations are positioned not as recipients of policing but as integral actors with their own contribution to recognition, victim support and platform engagement, a positioning that echoes the OSCE/ODIHR resource guide on the role of NGOs in preventing and responding to hate crimes.32
7. Investigation: standardisation, common language and the training rationale
The Investigation pillar absorbs the criminal-procedure dimension and integrates the training imperatives identified in the EXODUS public training needs analysis.9
7.1 From operational handling to judicial outcome
Investigation is the phase in which operational intelligence becomes judicially usable evidence. The framework addresses the legal characterisation of antisemitically-motivated offences under Greek and EU law, the role of prosecutors and examining magistrates, the evidentiary requirements for proving bias motivation, and the procedural safeguards that protect both the integrity of the evidence and the rights of the accused. The judicial dimension is particularly demanding for hate crimes, where motive is constitutive of the offence and must be proved to the appropriate evidentiary standard.29
7.2 Digital evidence and cross-border cooperation
Investigation in dual-environment cases is technically demanding. It involves digital evidence collection from platforms operating under different jurisdictions, requests through formal cross-border channels (mutual legal assistance, Europol/Interpol cooperation), and adherence to chain-of-custody requirements that hold up in court.21,33 The framework signals where established cybercrime investigative practice can be transferred to hate-crime cases.
7.3 Standardisation as an adult-learning imperative
The framework’s emphasis on standardised TTPs and a common operational vocabulary is the direct expression of the training-needs analysis.9 Standardisation here is not about rigidity; it is about portability of skills, of evidence, of procedure across investigators, between agencies, and over time. A first responder trained in the framework should produce field documentation that an investigator in another district can use without re-interpretation, and that a prosecutor can rely on without rebuilding the bias case from scratch.
This emphasis has substantive grounding in adult-learning theory. Kolb’s experiential learning cycle34 (concrete experience, reflective observation, abstract conceptualisation, active experimentation) models the way practitioner knowledge is acquired and consolidated; the framework’s combination of scenario-based learning, reflective protocols and standardised procedures maps directly onto this cycle. Mezirow’s transformative learning theory35 adds a further layer: encountering antisemitic phenomena that do not fit a practitioner’s prior frame of reference is, in his terms, a “disorienting dilemma” that requires critical reflection rather than rote application of rules. The framework’s emphasis on contextual literacy, rather than checklist application of indicators, is consistent with this view. Illeris’s integrated model of adult learning36 underscores that competency formation requires the simultaneous engagement of cognitive, emotional and social dimensions; victim-sensitive practice, in particular, cannot be reduced to a procedural protocol without the affective and relational competences that surround it. Recent analytical work on the design of practitioner training37 further supports the framework’s investment in common language as a precondition for transferable practice.
7.4 Alignment with the OSCE/ODIHR TAHCLE tradition
The framework’s training rationale stands in deliberate continuity with the OSCE/ODIHR Training Against Hate Crime for Law Enforcement (TAHCLE) programme.38 TAHCLE has, since its development in the early 2010s, provided the de facto European reference for LEA training on bias-motivated crime, with an experiential, scenario-based pedagogy that operationalises, often without naming it as such, the Kolb learning cycle. The framework proposed here builds on this tradition in three respects: it specialises the generic bias-crime logic for the antisemitism case (where symbolic and historical depth matters acutely); it extends the scope from physical to dual physical–digital environments (where TAHCLE in its core form is comparatively thinner); and it embeds the training rationale in a framework that holds together conceptual, operational and competency layers in a single instrument. The relationship is therefore one of inheritance and extension rather than substitution.
8. The cross-cutting victim-sensitive dimension
Running across the three pillars is an explicit victim-sensitive dimension, consistent with the EU Victims’ Rights Directive.39 The framework treats secondary victimisation, the additional harm caused by criminal-justice processes themselves, as a design constraint, not an aspirational add-on.
Concretely, this means: victim-centred interviewing techniques that avoid retraumatisation; trauma-informed handling of testimony; protocols that limit the number of times a victim must recount events; communication that keeps the victim informed of the procedural path; and explicit attention to the historical and collective dimensions of antisemitic victimisation, in which a single incident reactivates community memory of past persecution.40 The framework also addresses the chilling effect of online antisemitism on Jewish public presence, which has been documented across European contexts.41,42 The framework’s victim-sensitive orientation aligns with OSCE/ODIHR guidance on understanding the needs of hate crime victims43 and with the OSCE/ODIHR Practical Guide’s specific attention to the security needs of Jewish communities.25
The victim-sensitive dimension is methodologically significant: it forces the framework to evaluate procedural choices not only by their evidential yield but by their impact on the people the framework is meant to protect. This dual optimization, evidentially robust and victim-protective, is one of the framework’s defining commitments.
9. Discussion
Three observations follow from the analysis. First, the framework’s distinctiveness lies in integrating three layers that typically remain separate in the hate-crime literature: conceptual recognition, operational coordination and competency development. Hate-crime research has generated substantial knowledge in each of these domains,44 but the practical synthesis required by a working response chain has been comparatively underdeveloped. The framework proposed here offers one such synthesis, anchored in a specific legal-procedural context (Greek law) and aligned with established international reference frameworks (OSCE/ODIHR, EU Digital Services regime).
Second, the framework is grounded in practice rather than imposed on it. Its provisions are traceable, through the EXODUS project’s public analytical work,8,9,10 to evidence about actual conceptual ambiguities, actual operational frictions and actual training deficits. The adult-learning theorisation in Section 7.3 reframes this empirical grounding in terms that connect the framework to established scholarship on practitioner training, lending its standardisation emphasis a principled rather than merely pragmatic basis.
Third, the framework’s logic is transferable, though not automatically. The 3i structure, the whole-chain cooperation principle, the physical–digital integration and the victim-sensitive dimension are not specific to antisemitism. They could, in principle, structure responses to other forms of bias-motivated offending, anti-Muslim hatred, anti-LGBTQ+ violence, racist offending, disability hate crime, provided the underlying conceptual, operational and training analyses are redone for each victim group and legal-operational context. What is transferable is the architecture, not the content.
Limitations should be acknowledged. The article is a synthetic essay rather than an empirical evaluation: it does not measure whether the framework changes practitioner behaviour, whether it improves case outcomes, or whether it reduces secondary victimisation. Those evaluations would require longitudinal data collection beyond the scope of this article and remain a priority for future research. The framework’s effectiveness will ultimately be judged by its uptake, by adaptation across legal systems, and by measurable change in the conditions that allow antisemitic hate crimes to occur (the third “i” of Ratcliffe’s original triad).12 A further limitation is that several elements of operational practice within the EXODUS project remain, appropriately, outside the public domain, so the discussion here addresses operational principles rather than tactical detail.
10. Conclusions
This article has proposed a coherent operational framework for addressing antisemitic hate crimes in physical and digital environments. The framework’s principal contribution is not the invention of new procedures but the integration of conceptual, operational and training-informed foundations into a single 3i-based architecture, anchored in Greek law and the OSCE/ODIHR–EU regulatory environment, oriented to whole-chain cooperation, and grounded in a victim-sensitive logic supported by mainstream adult-learning theory. The framework’s architecture is generalisable beyond antisemitism, while its content must be re-grounded for each victim group and legal context. As bias-motivated offending continues to evolve at the intersection of physical and digital environments, frameworks of this kind, practice-informed, intelligence-led, victim-centred and educationally coherent, offer one of the more promising directions for European criminal-justice practice.
Ethics and consent
Ethical approval and consent were not required for this study. The research did not involve human participants, personal data, or real-world operational datasets. The article is based exclusively on the analysis of publicly available materials and open academic literature.
Use of generative AI tools
The authors used a generative AI tool (Claude, by Anthropic; Claude Opus 4.6 family) to support the streamlining of an earlier draft, the structuring of sections, and editorial refinement of language. The tool was not used to generate original scientific content, to interpret results, or to draw conclusions. All content was reviewed, verified and edited by the authors, who take full responsibility for the article’s accuracy and integrity.
Data availability
Data sharing is not applicable to this article as no data were created or analysed in this research.
The article is a synthetic essay that draws exclusively on publicly available sources cited in the References section. No primary data surveys, interviews, observations, experimental results, datasets or computational outputs were generated, collected or analysed at any stage. Consequently, no data repository deposit, dataset DOI or accession number is applicable. All sources used are openly accessible to readers and reviewers through the citations provided.
Acknowledgements
The authors would like to thank all partners of the EXODUS project consortium and, in particular, the Hellenic Police as project Coordinator, the Center for Security Studies (KEMEA) and the Institute for the Prevention and Treatment of Violence and the Promotion of Gender Equality, for their collaboration on the analytical work that underpins this article. The authors also gratefully acknowledge the contribution of the Greek Ministry of Justice, the Athens Lifelong Learning Institute (ALLI), the Jewish Community of Thessaloniki, and the Greek National Commission for Human Rights (NCHR) to the implementation of the EXODUS project. The authors are further grateful to colleagues at KEMEA and the Hellenic Police for their valuable feedback during the preparation of this manuscript.
References
- 1.
Becker MJ, Ascone L, Troschke H:
Antisemitic comments on Facebook pages of leading British, French, and German media outlets.
Humanities and Social Sciences Communications.
2022; 9(1). PubMed Abstract
| Publisher Full Text
| Free Full Text
- 2.
Santus D, Bettin CM:
Antisemitism in the algorithmic age: propaganda, disinformation, and generative AI. Zeitschrift für Religion, Gesellschaft und Politik.2025. Publisher Full Text
- 3.
De Smedt T:
Online antisemitism across platforms. arXiv.2022. Publisher Full Text
- 4.
Lupu Y, Sear R, Velásquez N, et al.:
Offline events and online hate.
PLOS ONE.
2023; 18: e0278511. PubMed Abstract
| Publisher Full Text
| Free Full Text
- 5.
Williams ML, Burnap P, Javed A, et al.:
Hate in the Machine: Anti-Black and Anti-Muslim Social Media Posts as Predictors of Offline Racially and Religiously Aggravated Crime.
The British Journal of Criminology.
2020; 60(1): 93. Publisher Full Text
- 6.
OSCE/ODIHR:
Hate Crime Data-Collection and Monitoring Mechanisms: A Practical Guide.
Warsaw:
OSCE/ODIHR;
2014.
- 7.
Hagerlid M, Granström G:
Hate Crime Investigation and Sentencing in Sweden: What Have We Learned in the Past 20 Years?.
Eur. J. Crim. Policy Res.
2023; 31: 193–210. Publisher Full Text
- 8.
Center for Security Studies:
Conceptual and Contextual Framework on Antisemitism (Deliverable D2.1). Public deliverable, EXODUS project, CERV Grant Agreement No. 101144655.2025.
Reference Source
- 9.
Athens Lifelong Learning Institute (ALLI):
Training Needs Analysis (EXODUS Deliverable D2.3). EXODUS project, CERV Grant Agreement No. 101144655.2025.
Reference Source
- 10.
Center for Security Studies:
Handbook on Antisemitic Symbols and Recognition (Deliverable D3.1). Public deliverable, EXODUS project, CERV Grant Agreement No. 101144655.2025.
Reference Source
- 11.
Gale NK, Heath G, Cameron E, et al.:
Using the framework method for the analysis of qualitative data in multi-disciplinary health research.
BMC Med. Res. Methodol.
2013; 13: 117. PubMed Abstract
| Publisher Full Text
| Free Full Text
- 12.
Ratcliffe JH:
Intelligence-Led Policing.
Routledge;
2nd ed2016. Publisher Full Text
- 13.
OSCE Transnational Threats Department/Strategic Police Matters Unit (TNTD/SPMU):
Intelligence-Led Policing: A Guidebook (OSCE Publication Series, Vol. 13).
Vienna:
OSCE;
2017b.
- 14.
Greek Penal Code:
Law 4619/2019, Article 81A - Crime committed for reasons related to race, colour, religion, descent, national or ethnic origin, sexual orientation, gender identity or disability of the victim (aggravating circumstance).2019.
- 15.
Enstad JD:
Accounting for Contemporary Antisemitism: A Four-Dimensional Framework and a New Dataset.
Antisemitism Studies.
2023; 7(1): 4–49. Publisher Full Text
- 16.
Sharansky N:
3D Test of Anti-Semitism: Demonization, Double Standards, Delegitimization.
Jewish Political Studies Review.
2004; 16(3–4).
- 17.
OSCE/ODIHR:
Hate Crime Prosecution at the Intersection of Hate Crime and Criminalised Hate Speech.
Warsaw:
OSCE/ODIHR;
2024.
- 18.
IHRA (International Holocaust Remembrance Alliance):
Working Definition of Antisemitism. Adopted Plenary, Bucharest, 26 May 2016.2016.
- 19.
Jikeli G, Karali S, Miehling D, et al.:
Antisemitic Messages? A Guide to High-Quality Annotation and a Labeled Dataset of Tweets.
arXiv.
2023. Publisher Full Text
- 20.
Mihaljević H, Steffen E:
How toxic is antisemitism? Potentials and limitations of automated toxicity scoring for antisemitic online content.
arXiv.
2023. Publisher Full Text
- 21.
Olber P:
The survey on cross-border collection of digital evidence by representatives from Polish prosecutors' offices and judicial authorities.
J. Digit. Forensic Secur. Law.
2021; 16(2). Publisher Full Text
- 22.
Pezzella FS, Fetzer MD, Keller T:
The Dark Figure of Hate Crime Underreporting.
Am. Behav. Sci.
2019; 000276421882384. Publisher Full Text
- 23.
Krupa JM, Holt TJ, Fisher T, et al.:
Cybercrime and Cyber Investigations: Development of a Training Curriculum for Law Enforcement Personnel.
Journal of Criminal Justice Education.
2025; 37: 232–255. Publisher Full Text
- 24.
OSCE/ODIHR:
Hate Crime Laws: A Practical Guide.
Warsaw:
OSCE/ODIHR;
Second ed2017a.
- 25.
OSCE/ODIHR:
Practical Guide on Understanding Antisemitic Hate Crimes and Addressing the Security Needs of Jewish Communities.
Warsaw:
OSCE/ODIHR.
2018.
- 26.
Chandra M, Pailla DR, Bhatia H, et al.:
"Subverting the Jewtocracy": Online Antisemitism Detection Using Multimodal Deep Learning. WWW '21.2021. Publisher Full Text
- 27.
Liu D, Wang M, Catlin AG:
Detecting Anti-Semitic Hate Speech using Transformer-based Large Language Models.
arXiv.
2024. Publisher Full Text
- 28.
Ramponi A, Testa B, Tonelli S, et al.:
Addressing religious hate online: from taxonomy creation to automated detection.
PeerJ Computer Science.
2022; 8: e1128. PubMed Abstract
| Publisher Full Text
| Free Full Text
- 29.
Bilius M, Navickienė Ž, Velička V:
Hate Crimes: Evaluation of Lithuanian Courts' Decisions in the Light of the Practice of the European Court of Human Rights.
Baltic Journal of Law & Politics.
2021; 14(1): 22–47. Publisher Full Text
- 30.
European Union:
Regulation (EU) 2022/2065 of the European Parliament and of the Council on a Single Market For Digital Services (Digital Services Act).2022.
- 31.
European Commission:
EU Code of Conduct on countering illegal hate speech online plus (revised).
Brussels:
European Commission;
2025.
- 32.
OSCE/ODIHR:
Preventing and Responding to Hate Crimes: A Resource Guide for NGOs in the OSCE Region.
Warsaw:
OSCE/ODIHR;
2009.
- 33.
Mutawa NA, Bryce J, Franqueira VNL, et al.:
Behavioural Digital Forensics Model: Embedding Behavioural Evidence Analysis into the Investigation of Digital Crimes.
Digit. Investig.
2018; 28: 70–82. Publisher Full Text
- 34.
Kolb DA:
Experiential Learning: Experience as the Source of Learning and Development.
Pearson FT Press;
2nd ed2015.
- 35.
Mezirow J:
Learning to think like an adult: Core concepts of transformation theory.Taylor EW, Cranton P, editors.
The Handbook of Transformative Learning: Theory, Research, and Practice.
Jossey-Bass;
2012; pp. 73–95.
- 36.
Illeris K:
Transformative learning re-defined: as changes in elements of identity.
Int. J. Lifelong Educ.
2014; 33(5): 573–586. Publisher Full Text
- 37.
Hoggan C:
Analytical tools for adult learning research.
Foundations of Adult and Continuing Education.
Wiley;
2022.
- 38.
OSCE/ODIHR:
Training Against Hate Crime for Law Enforcement (TAHCLE) — Trainer's Manual (December 2025 edition).
Warsaw:
OSCE/ODIHR;
2025.
- 39.
European Union:
Directive 2012/29/EU of the European Parliament and of the Council establishing minimum standards on the rights, support and protection of victims of crime.2012.
- 40.
Fireberg H:
A "Jewish Problem" or a "Society Problem"? Understanding Contemporary Antisemitism in Europe from Jewish and Governmental Perspectives.
De Gruyter eBooks;
2023; 87. Publisher Full Text
- 41.
Khaleghipour M, Koban K, Matthes J:
Listen to Me! Target Perceptions of Digital Hate: A Scoping Review of Recent Research.
Trauma Violence Abuse.
2025; 26(5): 1082–1096. PubMed Abstract
| Publisher Full Text
| Free Full Text
- 42.
Matthes J, Koban K, Buehrer S, et al.:
The State of Evidence in Digital Hate Research: An Umbrella Review.
Commun. Res.
2025. Publisher Full Text
- 43.
OSCE/ODIHR:
Understanding the Needs of Hate Crime Victims.
Warsaw:
OSCE/ODIHR;
2020.
- 44.
Vergani M, Perry B, Freilich JD, et al.:
Mapping the scientific knowledge and approaches to defining and measuring hate crime, hate speech, and hate incidents: A systematic review.
Campbell Syst. Rev.
2024; 20: e1397. PubMed Abstract
| Publisher Full Text
| Free Full Text
