October 8, 2026 — Warsaw, Poland
Recommendations of American Jewish Committee
Submitted by Holly Huffnagle, Director of Antisemitism Policy
to the Organization for Security and Cooperation in Europe (OSCE)
Warsaw Human Dimension Conference
Plenary Session 4: October 8, 2026
Warsaw, Poland
Preventing Antisemitism Is Essential to Europe’s Security:
Meeting the OSCE’s Responsibility in a New Threat Environment
In March 2026, in suburban Detroit, a man drove a truck loaded with gasoline and fireworks into Temple Israel, one of the largest synagogues in the United States. Children were inside its preschool. He exchanged gunfire with synagogue security before dying at the scene. The FBI called it a “targeted act of violence against the Jewish community.”[1] In the days that followed, a disturbing context emerged. The attacker, who had been born in Lebanon, had recently lost close family members in an Israeli airstrike there. Public reporting linked his deterioration to those deaths, and U.S. federal investigators stated he was motivated and inspired by Hezbollah’s militant ideology.[2] But his choice of target should trouble us all: not a government building or military installation, but a synagogue thousands of miles away, where Jewish families and young children were simply living their lives.
I begin with this American tragedy because it also reveals a European reality: conflicts and propaganda originating beyond a country’s borders can become threats to people within them. American Jewish Committee, the global advocacy organization for the Jewish people, sees this challenge firsthand through 40 offices and dozens of partnerships with Jewish communities worldwide. Across markedly different political and security environments, the warning is consistent: transnational threats can reach Jewish communities through both physical violence and the information environment, affecting whether Jews can gather, worship, study, and live openly and safely. But the danger does not stop at the doors of Jewish institutions. The same forces that exploit antisemitic conspiracy, scapegoating, extremist mobilization, and manipulated information also erode public trust, fracture democratic societies, and threaten the security of all. Europe has already seen how quickly that danger can become concrete.
In August 2024, a man carrying a Palestinian flag attempted to set fire to the synagogue in La Grande-Motte, France, while worshippers were present. He ignited vehicles outside, causing the explosion of a nearby gas cylinder. French authorities treated the attack as terrorism and said a catastrophe had narrowly been avoided.[3] In February 2025, four alleged Hamas members went on trial in Germany, accused of preparing attacks on Jewish institutions in Europe.[4]
This year, twenty-four governments—including many represented in this room—gave unusually clear collective recognition to this transnational threat, condemning lethal plotting and other malign activity linked to Iranian security services against Jewish and Israeli communities and other targets across Europe, North America, and Australia. The United Kingdom later assessed that members of Iran’s Islamic Revolutionary Guard Corps Quds Force “almost certainly directed” attacks across Europe by Harakat Ashab al-Yamin al-Islamiya (HAYI), also known as the Islamic Movement of the Companions of the Right (IMCR). The group publicly claimed attacks on Jewish-linked locations, including the burning of four Hatzolah ambulances in London.[5]
These incidents differ in important ways. One involves a lone attacker; others involve a terrorist organization or the machinery of a foreign state. We should not collapse them into a single phenomenon. Yet each exposes the same vulnerability: antisemitic actors operating across borders can invoke foreign conflicts as a pretext for targeting Jews. In doing so, they give antisemitic violence and intimidation a political guise while advancing broader extremist and anti-democratic aims that spread fear, division, and instability throughout society.
Antisemitism is not only hatred and discrimination; it is also a security vulnerability. It is built on conspiracy, distrust, scapegoating, and the myth of hidden power—the very ingredients hostile actors use to undermine democratic institutions, fracture societies, manipulate information environments, and threaten the security of everyone, not only Jews.
The danger is not new. What is new is the speed, scale, and sophistication with which it can be weaponized. The argument advanced here is therefore not a departure from the OSCE’s understanding of security, but a return to it.
The OSCE was built on the recognition that military security alone is not enough. Lasting security can only be achieved through respect for human rights, strong democratic institutions, tolerance, and the rule of law.[6] The 1992 Helsinki Document made this explicit, describing the protection of human rights and the strengthening of democratic institutions as a “vital basis” for comprehensive security.[7]
In 2004, participating States meeting in Berlin acknowledged that antisemitism had assumed new forms and posed a threat to security and stability across the OSCE region. They committed to practical action, including better hate-crime data, stronger legal protections, and expanded education. But the Berlin Declaration preceded a transformation in the information environment that now operates alongside those physical threats: the rise of social media, algorithmic recommendation systems, large-scale bot networks, coordinated inauthentic behavior, generative artificial intelligence, and AI systems capable of producing and adapting persuasive content across languages.
Today, an antisemitic narrative can be seeded or exploited in one country, amplified by coordinated networks in another, translated instantly into multiple languages, optimized by algorithms for engagement, and reach citizens across borders within minutes. Foreign intelligence services, terrorist organizations, extremist movements, criminal networks, and ordinary users now operate in the same information ecosystem.
This reality is blurring the old boundaries between foreign and domestic threats, physical and cyber security, and hate crime and national security. When antisemitism is used to divide societies, corrode trust, radicalize individuals, and create openings for hostile actors, preventing it can no longer be the responsibility of minority-affairs or anti-discrimination offices alone.
So the question before us in 2026 is not whether antisemitism is a security issue. The OSCE has already answered that question. The question is whether our security architecture has kept pace with the way the threat now operates.
The new security environment
The contemporary threat spans physical and digital environments. Hostile actors can exploit antisemitism offline through intimidation, proxy violence, and attacks on Jewish targets, and online through coordinated operations that weaponize antisemitic narratives and increasingly use AI to scale them. These threats operate through three distinct but related mechanisms: transnational coercion and proxy violence, antisemitic narrative infrastructure for influence operations, and AI as a force multiplier for antisemitic influence.
- Antisemitism as a vehicle for transnational coercion and proxy violence
First, foreign states and their proxies are projecting overseas conflicts into OSCE participating States by targeting Jewish communities, institutions, and individuals. These actors are not simply conflating Jews with Israel; they deliberately invoke conflicts involving Israel to legitimize antisemitic targeting.
The evidence is concrete. The joint condemnation described above placed these attacks within a wider pattern of Iranian-linked coercion against dissidents, journalists, and Jewish and Israeli targets across several regions. A July 2025 USCIRF country update similarly documented Iranian-linked efforts to assassinate a rabbi in Azerbaijan, recruit criminal gangs to target Israeli embassies and Jewish institutions in Europe, and identify possible Jewish targets in Germany.[8] Violence and intimidation against Jews then become instruments of transnational coercion and repression, violating state sovereignty and the capacity of democratic governments to protect people on their territory. - Antisemitism as a narrative infrastructure for influence operations
Second, antisemitism provides a ready-made narrative architecture for covert influence operations. Claims of secret power, captured institutions, financial control, manipulated wars, corrupt media, and dual loyalty allow malign networks to transform contemporary controversies into evidence of generalized betrayal. The security concern is not criticism of Israel, however harsh, but the covert manipulation of antisemitic narratives to interfere in domestic debate or facilitate radicalization.
This activity falls within the broader security challenge that the European Union describes as Foreign Information Manipulation and Interference, or FIMI: intentional, coordinated efforts by state or non-state actors, including proxies, to manipulate the information environment and undermine democratic trust.[9] Antisemitism is especially useful in such operations because it supplies ready-made narratives of hidden power, betrayal, conspiracy, and blame.
A 2026 ShadowGraph Intelligence investigation identified an Iran-aligned network of at least 21 ostensibly independent media accounts on X that generated roughly 137,000 posts and 48.6 million engagements over six months. The accounts amplified one another, used fabricated quotations and pseudo-news branding, and advanced a consistent pro-Iran, anti-U.S., and anti-Israel line.[10] The investigation does not establish that every message was antisemitic or identify a state sponsor, but it demonstrates how coordinated networks can manufacture apparent consensus, outrage, popularity, or momentum and conceal foreign intervention in domestic political debate. - AI as a force multiplier for antisemitic influence
Third, generative and conversational AI are changing the economics and structure of antisemitic influence. They make content cheaper to produce, easier to adapt across languages and formats, and increasingly capable of personalized delivery. AI can make antisemitism more dangerous by reproducing familiar conspiracies and dehumanizing tropes at unprecedented speed and scale. When used for incitement, radicalization, recruitment, or coordinated foreign influence, that capacity is a security concern.
Five Responsibilities for OSCE Participating States
If hostile actors are exploiting antisemitism as a security vulnerability, declarations of concern are not enough. Participating States should translate existing OSCE commitments into five practical responsibilities that protect Jewish communities and strengthen democratic resilience and collective security across the region.
- Establish a national coordination mechanism
Each participating State should designate a clearly accountable senior body or official to coordinate foreign affairs, interior, justice, intelligence, law enforcement, cyber, education, equality bodies, and Jewish communities. Whatever its institutional form, it should maintain a shared threat picture, assign responsibilities, identify gaps, and monitor implementation.
AJC’s Combating Antisemitism Playbook, developed with more than 60 practitioners and experts across 13 sectors, can help governments identify unassigned responsibilities, information-sharing gaps, and sectors not yet engaged.[13] Although developed for the United States, its central principle applies broadly: government coordination organizes the response, while sustained civil-society participation builds resilience. - Disrupt the infrastructure of foreign coercion and coordinated manipulation
Participating States should build cross-government capacity to identify and disrupt the infrastructure used for foreign coercion and manipulation, including criminal proxies, financing and logistical networks, bot farms, influence-for-hire services, account-farming operations, and coordinated inauthentic networks. Authorities should share indicators and use appropriate criminal, cyber, sanctions, fraud, and anti-money-laundering tools against the actors and services behind these operations. Responses to information operations should focus on deceptive conduct—such as concealed sponsorship, fabricated identities, automation, and fraudulent engagement—not political viewpoint. Participating States should integrate antisemitic narratives and networks into broader FIMI threat analysis, rather than treating online antisemitism and foreign information manipulation as separate policy silos.
ODIHR and the Transnational Threats Department should establish a standing cross-dimensional mechanism to develop common terminology, content-neutral indicators, attribution standards, cross-border warning procedures, and rights-respecting guidance for disrupting deceptive networks.[14] - Require greater transparency about the origin and authenticity of online influence
Participating States should require major platforms to provide meaningful transparency about the provenance and authenticity of online influence, including reliable country-of-origin information, coordinated foreign amplification, unusually large or coordinated engagement, and enforcement against inauthentic networks.[15] Those requirements should include adequate linguistic and cultural expertise in every major market a platform serves.[16] For EU Member States, implementation of the Digital Services Act should ensure that coordinated foreign manipulation is meaningfully captured in systemic-risk assessments.[17] The aim is not to restrict cross-border participation or political viewpoints, but to help citizens distinguish authentic debate from manufactured consensus.
CyberWell’s research illustrates the present scale of the problem. Between January 2025 and February 2026, it verified more than 300 pieces of AI-generated antisemitic content across major platforms; 98.4 percent appeared after June 2025 and collectively generated more than 30 million views and 2.8 million interactions. Three-quarters used stereotypical or conspiratorial depictions of Jews, and one-third called for, glorified, or justified violence.[11] While these findings do not establish effects on attitudes or behavior, they demonstrate the scale and reach of AI-generated antisemitic content and how easily familiar antisemitic tropes can be converted into images, songs, fabricated news segments, and videos and distributed to mass audiences.
The Institute for Strategic Dialogue also warns that conversational AI can create “closed loops” in which users explore grievances and extremist ideas through private, continuous, personalized interaction.[12] Systems can adapt to a user’s language and concerns and remain available at any hour. The full security impact remains uncertain, but extremist ideas can now be explored not only in public feeds but through individualized exchanges. Because antisemitism supplies a ready vocabulary of hidden power, betrayal, corruption, and blame, it is particularly adaptable to this environment.
Taken together, these mechanisms show how antisemitism can be exploited across physical and information environments. This does not mean every antisemitic incident is foreign-directed or every antisemitic expression constitutes a security threat. It means malign networks and organizations can exploit a preexisting vulnerability through coercion, coordinated manipulation, and AI-enabled influence—and policy must distinguish among those mechanisms while accounting for how they interact. Preventing antisemitism is therefore not peripheral to European security; it is essential to constraining foreign coercion, protecting democratic debate, and strengthening societal resilience against manipulation. - Build security safeguards into AI systems
Participating States should incorporate antisemitic and extremist exploitation into national AI-security frameworks. For EU Member States, the AI Act provides an important regulatory foundation, particularly for general-purpose AI models with systemic risk. Building on that framework, AI providers should conduct multilingual safety testing, detect coordinated abuse, disclose significant safety failures, enable independent research, and maintain effective reporting channels for governments and civil society.
Governments should develop corresponding expertise across relevant agencies: foreign ministries on AI-enabled influence operations; intelligence and law-enforcement services on synthetic identities, deepfakes, and automated persuasion; cyber authorities on AI agents and automated networks; and officials combating antisemitism on the capacity of generative systems to reproduce and scale conspiratorial narratives. Policy should not rely solely on detecting harmful outputs after they appear; it should make manipulation harder to scale, easier to detect, and less able to masquerade as authentic human activity.[18] - Treat antisemitism as an early-warning indicator and share the warning across borders
Participating States should incorporate significant changes in antisemitic threats, incidents, and coordinated narratives into national early-warning and threat-assessment systems, while continuing to address antisemitism as a hate-crime, discrimination, and human-rights concern. Relevant indicators should be shared across ministries and assessed alongside evidence of extremist mobilization, foreign influence, criminal proxies, cyber activity, and preparation for violence.
The OSCE should facilitate rapid, rights-respecting exchange of indicators, network information, and threat assessments across participating States. A network detected in one country may have infrastructure, audiences, or targets in several others, while narratives can move across languages within hours. By connecting fragments of the same cross-border threat, the OSCE can support earlier detection, stronger attribution, and coordinated action.
The “Warsaw Test”: Five questions for Participating States
The recommendations above are ambitious, but they can be distilled into a practical test. Each participating State should be able to answer five questions.
- Who in our government sees the whole picture? Is there one mechanism capable of connecting antisemitic hate crime, threats to Jewish institutions, foreign influence, extremist mobilization, intelligence reporting, cyber threats, online manipulation, and the concerns raised by Jewish communities themselves? If those signals remain separated by ministry or mandate, then government may possess the facts without understanding the threat.
- Could we detect and disrupt a foreign actor using criminal proxies or exploiting antisemitic and other divisive narratives to manipulate domestic debate? Do we have the technical capacity, legal authorities, and interagency procedures to identify criminal proxies and deceptive networks, trace their financing and infrastructure, attribute sponsorship when the evidence supports it, and act before manufactured influence is normalized as authentic public opinion?
- Can our citizens tell whether online influence is authentic and where it originates? Do platforms provide meaningful transparency about the provenance of accounts and engagement, including bot amplification, coordinated manipulation, foreign engagement, and the geographic sources of viral activity—in the languages citizens actually use?
- Are we prepared for AI-enabled influence operations? Do government agencies have the expertise to assess generative AI, synthetic media, conversational systems, and AI agents used for manipulation or radicalization? Are providers testing for antisemitic and extremist exploitation across the languages of the OSCE region, detecting coordinated abuse, and maintaining effective government reporting channels when harms emerge?
- When antisemitism rises, do we treat it only as a Jewish-community concern—or as an early warning about the health and security of our societies? Do we assess those signals alongside foreign influence, extremist mobilization, cyber activity, and preparation for violence—and share relevant indicators across borders before fragments of the same threat become separate national crises?
The Warsaw Test is meant to prompt a different conversation inside government. The question is not only, “What are we doing to combat antisemitism?” It is also, “What does antisemitism reveal about vulnerabilities in our security environment—and are we organized to act before they become crises?”
Conclusion
Jews should not have to prove that protecting them benefits everyone else. Their safety, dignity, religious freedom, and ability to live openly as Jews are obligations in themselves and part of the human rights and fundamental freedoms that participating States have already committed to uphold.
But history teaches a broader lesson: antisemitism does not remain contained within the Jewish community. The conspiracies and scapegoating that target Jews also corrode institutional trust, legitimize political violence, and create openings for extremists, criminal networks, terrorist organizations, and hostile states. The task is not to securitize Jewish life, but to recognize antisemitism as both a grave harm and an early warning.
For the OSCE, this is the contemporary meaning of comprehensive and indivisible security. Human rights, democratic institutions, resilient information environments, and the security of States rise or fall together. Protecting Jewish communities and strengthening democratic resilience are not competing objectives; they are mutually reinforcing responsibilities.
Governments can make their societies harder to manipulate. They can connect what they know, disrupt the infrastructure of coercion and deception, demand transparency from the technologies shaping public discourse, prepare for AI-enabled threats, and act on warning signs before they become crises.
Preventing antisemitism protects Jews—and it protects us all. It defends the democracy, rule of law, and collective security on which every society depends.
1. Federal Bureau of Investigation, “Remarks by Jennifer Runyan, Special Agent in Charge of FBI Detroit, on Temple Israel Synagogue Attack,” March 12, 2026, https://www.fbi.gov/contact-us/field-offices/detroit/news/remarks-by-fbi-detroit-on-temple-israel-synagogue-attack.
2. Federal Bureau of Investigation, “Detroit Special Agent in Charge Jennifer Runyan Delivers Final Update on West Bloomfield Attack,” March 30, 2026, https://www.fbi.gov/contact-us/field-offices/detroit/news/detroit-special-agent-in-charge-jennifer-runyan-delivers-final-update-on-west-bloomfield-attack.
3. Reuters, “French Police Arrest Man Suspected of Attempted Arson Against Synagogue,” August 24, 2024, https://www.reuters.com/world/french-police-officer-hurt-suspected-arson-attack-front-synagogue-2024-08-24/.
4. Reuters, “Hamas Members Suspected of Plotting Attacks Go on Trial in Germany,” February 25, 2025, https://www.reuters.com/world/europe/hamas-members-suspected-plotting-attacks-go-trial-germany-2025-02-25/.
5. UK Home Office, “National Security (State Threats) Act 2026: Factsheet,” July 17, 2026, https://www.gov.uk/government/publications/national-security-state-threats-act-2026-factsheet/national-security-state-threats-act-2026-factsheet.
6. Organization for Security and Co-operation in Europe, “Our Mandate on Tolerance and Non-Discrimination,” accessed October 6, 2026, https://www.osce.org/node/660075.
7. OSCE Office for Democratic Institutions and Human Rights, OSCE Human Dimension Commitments, Volume 2: Chronological Compilation, 2nd ed. (Warsaw: OSCE/ODIHR, 2005), 129, https://odihr.osce.org/sites/default/files/f/documents/0/f/16378.pdf.
8. U.S. Commission on International Religious Freedom, Iran Country Update (Washington, DC: USCIRF, July 2025), 6, https://www.uscirf.gov/sites/default/files/2025-07/2025%20Iran%20Country%20Update.pdf.
9. European External Action Service, “Information Integrity and Countering Foreign Information Manipulation & Interference (FIMI),” March 17, 2026, https://www.eeas.europa.eu/eeas/information-integrity-and-countering-foreign-information-manipulation-interference-fimi_en.
10. Travis Hawley, “Anatomy of an Iran-Aligned Influence Network on X,” ShadowGraph Intelligence, July 19, 2026, https://shadowgraphintel.com/reports/iran-aligned-influence-network-on-x/.
11. CyberWell, AI-Generated Antisemitism: Abuse Trends and Safety Gaps Across Social Media and AI Platforms (2026), 6–7, 9, https://cyberwell.org/reports/ai-generated-antisemitism/.
12. Simeon Dukić, Natalie Vela, and Siddharth Venkataramakrishnan, Radicalisation in Closed Loops: Risks and Intervention Opportunities with AI Chatbots and Companions (London: Institute for Strategic Dialogue, July 22, 2026), 4, https://www.isdglobal.org/publication/radicalisation-in-closed-loops-risks-and-intervention-opportunities-with-ai-chatbots-and-companions/.
13. American Jewish Committee, AJC’s Combating Antisemitism Playbook: Strategies to Understand, Respond to, and Prevent Antisemitism (2026), https://www.ajc.org/playbook.
14. Organization for Security and Co-operation in Europe, Transnational Threats Department, “OSCE Transnational Threats Department,” accessed October 6, 2026, https://tntd.osce.org/.
15. Digital Borders, “Who Amplifies American Conversation, and From Where,” Digital Borders & Authenticity: Engager-Origin Study, July 2026, https://reports.byeredtriangles.com/geographic-analysis-of-amplifiers-across-accounts-keywords/.
16. American Jewish Committee and CyberWell, The State of Antisemitism in America: Findings and Recommendations for Major Digital Platforms (2026), 42, https://www.ajc.org/sites/default/files/pdf/2026-03/SOAR2025-Social-Media-Report.pdf.
17. Regulation (EU) 2022/2065 (Digital Services Act), arts. 34–35.
18. See also American Jewish Committee, “Strategies for Digital Platforms and Technology Companies,” AJC’s Combating Antisemitism Playbook (2026), https://www.ajc.org/playbook/tech-companies, which calls for stronger bot-network enforcement, reduced amplification of manipulated content, integrity in AI training data, multilingual enforcement, transparency, independent evaluation, and dedicated company leadership on antisemitism.
