Strategies for Digital Platforms and Technology Companies: Understanding, Responding to, and Preventing Antisemitism

Effective policies, moderation systems, and AI safeguards depend on recognizing how antisemitic narratives appear, spread, and adapt across society, platforms, and technologies.

• Recognize diverse sources of antisemitism—and its role as an early warning signal: Antisemitic narratives originate from multiple ideological ecosystems—including far-right and far-left political antisemitism, extremism in the name of Islam, conspiracy movements, and state-sponsored disinformation. These narratives frequently emerge at the forefront of broader online harms, including extremism and coordinated manipulation campaigns. Understanding these distinct drivers, including how antisemitic narratives often function as an early warning signal for broader harms, help platforms detect emerging threats earlier and strengthen overall trust and safety systems.

• Ensure protected-category frameworks accurately address Jews: Systems should recognize antisemitism that Jews as both a religion and a people, including coded, proxy, or indirect attacks. 
• Address the evolving and multimodal nature of antisemitism: Antisemitic narratives appear across text, images, memes, video, audio, gaming environments, and AI-generated content, often relying on humor, symbolism, coded language, or proxy terms to evade detection. These include euphemisms, dog whistles, and the use of substitute terms to target Jews indirectly or collectively. Because these forms evolve rapidly, detection tools, moderation systems, and policies must continuously adapt to identify and respond to emerging expressions of antisemitism.
• Understand how platform design can amplify antisemitism: Antisemitic narratives can spread rapidly through coordinated harassment campaigns, bot networks, coordinated disinformation campaigns, and algorithmic amplification. Recommendation systems, engagement optimization, and virality dynamics can unintentionally accelerate the normalization of antisemitic content if not carefully monitored and adjusted.
• Ensure integrity of AI training data: AI systems should be trained on reliable, high-quality data and evaluated for biases that may reproduce antisemitic stereotypes or conspiracies. Training inputs must be reviewed to prevent manipulation, coordinated distortion, or biased sources from shaping model outputs. 
• Provide contextual information and redirects to improve user understanding: When users encounter antisemitic conspiracy theories, Holocaust denial, or related narratives, platforms can provide contextual “redirects” to credible educational resources. These interventions—such as informational pop-ups or search redirects—can help disrupt misinformation and guide users toward accurate information.

When antisemitic content, harassment, or threats appear on digital platforms, technology companies must respond quickly, consistently, and at scale to protect users, disrupt abuse networks, and prevent harm both online and offline. Efforts to counter antisemitism must remain consistent with freedom of expression and legal protections. Not all offensive or controversial speech warrants removal; however, platforms still have a responsibility to address threats, coordinated abuse, incitement, manipulation, and content that violates their own policies.

• Strengthen policies and enforce consistently and at scale: Enforcement outcomes should feed into measurable improvements in detection systems and strengthen feedback loops with trusted researchers and civil society partners. 
  • Maintain clear, enforceable policies that cover Holocaust denial and distortion, conspiracy narratives, scapegoating, and harassment tactics such as false-flag or self-victimization claims. 
  • Apply policies consistently across all platforms, features, and AI systems. Ensure enforcement and detection systems are equally effective across languages and regions, not disproportionately concentrated in English-language content.
  • Reduce the visibility of harmful content through downranking and other friction measures (including limiting sharing, monetization, or virality) when removal is not appropriate. Platforms should also distinguish between lawful but harmful content, policy-violating content, and illegal conduct, applying responses proportionately.
  • Establish surge-response protocols for spikes in content during major events or crises, and strengthen moderation in high-risk spaces such as comment sections, live chats, and gaming environments. 
  • Combine automated detection with trained human review.  
• Detect early warning signals of violence: Detect early warning signals of violence: Invest in systems that identify indicators of potential real-world harm — including reconnaissance activity, threat signaling, or rhetoric that may warrant further assessment by relevant authorities — and integrate them into crisis-response protocols, including escalation to law enforcement when credible threats emerge.
• Disrupt coordinated abuse and repeat offenders: Antisemitic content often spreads through networks of repeat offenders, coordinated harassment campaigns, and cloned or duplicate accounts designed to evade enforcement:
  • Detect patterns of coordinated abuse and apply escalating penalties—including demonetization, suspension, or permanent removal. 
  • Remove accounts tied to sanctioned, extremist, or terrorist organizations and implement systems to detect attempts to bypass enforcement. 
  • Coordinate with other companies and trusted partners to identify cross-platform abuse networks and prevent coordinated campaigns from migrating between services.
• Improve reporting and user protection: Ensure users can easily report antisemitic content, including conspiracies and AI-generated hate, and that these reports feed directly into enforcement decisions and system improvements. Protect at-risk users from harassment, account compromise, and data misuse. 
• Improve transparency, correction, and appeals mechanisms: Strengthen tools that identify and correct antisemitic misinformation—including conspiracies blaming Jews for global events or unfounded claims of “false flag” operations. Use annotations, fact-checking tools, and contextual notices to counter manipulation while ensuring users receive accurate information. Maintain clear, accessible appeals processes so users can challenge moderation decisions and understand how those decisions are reviewed and resolved. Transparent appeals and review mechanisms help protect against inconsistent enforcement, political bias, or overreach. Finally, provide appropriate access for independent auditors and researchers to evaluate enforcement actions and improve system performance.

Preventing antisemitism on digital platforms requires proactive system design, transparent governance, and sustained investment in safety. Companies must anticipate emerging threats, address structural incentives that amplify harmful content, and work with experts and communities to strengthen resilience against evolving forms of online abuse.

• Strengthen policies and safety standards: Maintain clear, enforceable policies that explicitly address antisemitism, including Holocaust denial, conspiracy narratives, scapegoating of Jews, and evolving forms of harassment such as false-flag claims or self-victimization narratives. Apply these standards consistently across all platform features, generative AI systems, and emerging technologies.
• Reduce manipulation and inauthentic networks through structural safeguards: Coordinated campaigns, bots, and fake accounts can amplify antisemitic narratives and distort public discourse. Platforms should strengthen authenticity systems and redesign engagement mechanisms so that unverified or newly created accounts cannot immediately influence rankings, recommendations, or trending systems. 
  • Introduce graduated trust models that limit algorithmic impact until accounts demonstrate authentic behavior.
  • Require light verification or other friction before accounts can meaningfully influence engagement signals.
  • Ensure engagement metrics reflect genuine user activity rather than automated amplification.
  • Continuously evaluate incentives that make large-scale manipulation viable.
• Strengthen AI governance and data integrity: Design AI systems to prevent the generation or amplification of antisemitic content:
  • Conduct regular red-team testing against antisemitic prompts and conspiracies, ensure training data is not distorted by manipulated or biased sources, and provide transparency about model training inputs. 
  • Implement safeguards against synthetic media and deepfakes used to spread antisemitic disinformation or incite harassment.
  • Ensure AI companion systems and conversational agents include safeguards against grooming, extremist radicalization, antisemitic manipulation, and abusive dependency dynamics, particularly for minors.
  • Implement safeguards against synthetic media and deepfakes used to spread antisemitic disinformation or incite harassment.
• Align platform incentives with safety: Recommendation systems, engagement metrics, and monetization structures can unintentionally reward harmful content. Review and adjust these systems to reduce the amplification of antisemitic, abusive, or manipulative content, ensuring user safety over virality.
• Increase transparency and system accountability: 
  • Provide clear, consistent disclosure of the prevalence and visibility of antisemitism on the platform, how it is amplified, how it is detected, and policies enforced across languages and platforms. 
  • Publish meaningful transparency reports that include enforcement outcomes, recurrence rates, and the performance of AI tools. 
  • Allow independent experts to evaluate recommendation systems and algorithms to reduce reliance on opaque decision-making. 
  • Maintain regular engagement with affected communities—including Jewish organizations and antisemitism experts—to share updates on emerging risks, policy changes, and safety improvements.
• Increase geographic transparency and user awareness of content origin: Foreign networks and coordinated campaigns can manipulate public discourse by masking the origin, location, or authenticity of accounts, including by appearing as domestic voices. Platforms should provide clear, visible indicators of account location and engagement sources to help users assess authenticity and reduce the impact of foreign influence operations.
  • Display country-of-origin labels for accounts, posts, and trending topics.
  • Provide users with information about where engagement originates.
  • Allow users to control the geographic mix of content in their feeds for sensitive civic or political topics.
  • Downrank or flag unverifiable or masked location signals on high-risk content.
• Invest in trust, safety, and partnerships: Sustained prevention requires well-resourced trust and safety teams and strong collaboration with researchers, civil society organizations, and affected communities:
  • Maintain dedicated expertise on antisemitism and hate speech, including a designated internal lead responsible for antisemitism and Jewish community engagement. 
  • Regularly engage with Jewish communities and antisemitism experts to share updates on threats, policy changes, and enforcement outcomes. 
  • Provide meaningful data access for independent researchers, strengthen partnerships that identify emerging threats, and support digital literacy initiatives that help users recognize and resist antisemitic narratives and misinformation. 

Educational Resources

• AJC’s Translate Hate, a visual glossary of words and symbols that are antisemitic or can become antisemitic in context
• AJC’s State of Antisemitism in America Reports
• AJC’s Recognizing When Anti-Israel Actions Become Antisemitic
• AJC’s 10 Tough Questions on Antisemitism Explained
• CyberWell’s Self-Victimization in Antisemitic Discourse: Analysis of the Online Aftermath of Violent Attacks on Jews and Israelis
• CyberWell’s Antisemitism Online Amid National Elections (2024-2025)
• Center for Countering Digital Hate’s Killer Apps: How mainstream AI chatbots assist users planning violent attacks
• Anti-Defamation League’s Generating Hate: Anti-Jewish and Anti-Israel Bias in Leading Large Language Models
• For additional information, reach out to [email protected] 

Practical Resources

• AJC’s The State of Antisemitism in America: Findings and Recommendations for Major Digital Platforms
• Institute for Strategic Dialogue (ISD)’s Mainstreaming Digital Human Rights: A pan-European Policy Roadmap to Combat Online Antisemitism
• AJC’s Reporting Antisemitism to Federal Agencies, Law Enforcement, and Social Media and AI Companies: Guidance by American Jewish Committee
• The ADIR Challenge’s Innovate to Illuminate, and Ideation Competition
• ISD, B’nai B’rith International, and UNESCO’s Online Antisemitism: A Toolkit for Civil Society
• Antisemitism Policy Trust’s Find Your Meme: How Antisemitic Memes Go Mainstream
• Antisemitism Policy Trust’s Detecting Deepfakes: Artificial Intelligence and Anti-Jewish Hate: A Case for Regulating Generative AI