ISD (Institute for Strategic Dialogue)

The Islamist extremist attack targeting Berlin's Christopher Street Day celebrations has once again highlighted that the threat posed by Islamic State (IS) and its supporters remains both persistent and evolving. A new ISD dispatch examines what the attack tells us about today's Islamist extremist landscape. While the investigation remains ongoing, the analysis identifies five key lessons: 🔴 IS remains a major terrorist threat, continuing to inspire and attract a new generation of supporters despite the loss of its territorial caliphate. 🔴 LGBTQ+ events have become an increasingly prominent target for Islamist extremist attacks and plots. 🔴 The attack raises important questions about the authorities’ handling of the case and how sentencing, risk assessment and the use of deradicalisation programmes can be improved. 🔴 Online extremist ecosystems continue to fuel radicalisation, with terrorist propaganda and legal but extremist content remaining readily accessible across digital platforms. 🔴 Pro-IS online networks rapidly exploit attacks to glorify perpetrators, encourage copycat violence and spread propaganda, increasingly using AI-generated content and mainstream social media platforms. The research also raises important questions about prevention. It examines the alleged attacker's known history of radicalisation, previous attempts to join IS, engagement with deradicalisation programmes and contact with authorities, underscoring the importance of understanding where existing prevention and risk management approaches can be strengthened. Understanding what happens after an attack is just as important as understanding the attack itself. By analysing how extremist movements adapt, exploit digital platforms and seek to inspire further violence, researchers, policymakers, technology companies and practitioners can better identify emerging risks, strengthen prevention efforts and develop more effective responses to terrorist and extremist content online. Read the full analysis: https://lnkd.in/ed2FW7FJ

Today, ISD is launching the Authoritarian Interference Tracker, a redesigned tool mapping how Russia, China and Iran are interfering in democracies across Europe and North America. Originally launched by the Alliance for Securing Democracy, which merged with ISD this year, the tracker now documents more than 290 incidents since the start of 2022 across five areas: cyber operations, kinetic operations, information operations, malign finance and political and civic subversion. The tracker, which will be continuously updated, includes an updated interface and new features to support researchers, policymakers and journalists in identifying trends across countries, threat actors and incident types. Using the tracker, ISD analysts found that: 🔴 Russia remains the most prolific threat actor, linked to 188 documented incidents since 2022.  🔴 Iran is rapidly intensifying its activity, with 18 incidents recorded in the first half of 2026 alone.  🔴 Foreign actors are increasingly targeting energy and water systems through disruptive cyberattacks. One Russian GRU-linked cyberattack burst pipes and disrupted the water supply to hundreds of homes outside Copenhagen.  🔴 Russia and Iran are escalating on-the-ground operations, including sabotage, arson and assassination plots. Documented Russian operations increased from two in 2022 to nearly 60 since the start of 2024.  🔴 Both states increasingly rely on disposable agents and criminal networks recruited online and paid in cryptocurrency.  🔴 China’s most aggressive interference takes the form of transnational repression against diaspora communities and critics abroad, ranging from illegal overseas “police stations” to AI-generated explicit deepfakes of a Canada-based critic of Beijing. Explore the tracker here: https://lnkd.in/gr54Nqrm  

Conversational AI is transforming how people seek information, advice and companionship. But what happens when AI chatbots become part of the pathway to radicalisation? New ISD research examined approximately 1,800 responses from ChatGPT, Claude, Gemini, Grok, DeepSeek, Llama, Character.ai, Replika, Nomi and Gab.ai's Arya to assess how they respond to extremist and conspiratorial narratives. The findings reveal that AI chatbots and companions can create a ‘closed loop’, one-to-one environments where harmful beliefs are reinforced over time. Unlike social media, these conversations take place without social correction or outside challenge, creating new risks to existing online safety frameworks. Key findings from the report include: 🟥 Risk was driven more by model design than by user prompts. 🟥 Compared to general-purpose chatbots, AI companions were associated with higher risk, as they were more likely to validate harmful narratives through uncritical and relational responses. 🟥 Risk increased as conversations progressed, suggesting safeguards can weaken over longer interactions. 🟥 Higher-risk responses were far more likely to reinforce grievances and conspiracy theories than to explicitly promote violence. The research also shows that these outcomes are not inevitable. Some models consistently recognised signs of radicalisation, challenged harmful narratives and steered users towards safer forms of engagement, demonstrating that better design can reduce risk. As conversational AI becomes embedded in everyday life, safety cannot be measured by individual responses alone. Policymakers and AI developers must account for the cumulative impact of sustained, personalised, private and emotionally engaging conversations. Read the report: https://lnkd.in/e8xhB9aC

Gendered deepfakes, or synthetic non-consensual intimate images (NCII), are one of the fastest-growing aspects of technology-facilitated gender-based violence. While advances in generative AI have made NCII easier to create, the problem extends far beyond the technology itself. A complex ecosystem of websites, platforms, payment systems, infrastructure providers and commercial incentives enables users to discover, create, distribute and monetise gendered deepfakes at scale, while reinforcing broader patterns of online misogyny. Drawing on six months of research, ISD's new report maps the entire ecosystem of ‘nudify’ tools and identifies where governments, regulators and technology companies can intervene most effectively. It examines how deepfake content is created, how users are directed to these services, how operators generate revenue and the technical infrastructure that allows these tools to flourish despite growing legislative and regulatory attention. The research identified 181 publicly accessible nudify websites receiving more than 40 million unique visitors every month, with more than 5.7 million visits originating from mainstream social media platforms. It also found that many of these services remain easily discoverable through search engines and app stores, despite existing platform policies, highlighting the role that mainstream digital services continue to play in enabling this ecosystem. The report also finds that the creation and sharing of gendered deepfakes are often driven by more than sexual motives. Analysis of user behaviour and online discussions shows that this content is frequently used to humiliate, degrade and exert power and control over women and girls, reinforcing broader patterns of online misogyny and underscoring the need to treat this as a systemic form of technology-facilitated gender-based violence rather than simply a challenge posed by AI. Read the report here. https://lnkd.in/eDbDzQUw

📣 On 21 July: Join Us for Our Next Deep Dive: ISD Research on Digital Regulatory Implementation   This session presents recent research on the implementation of platform regulation, focusing on how regulatory provisions are applied in practice. It examines early findings on the implementation of risk identification and mitigation measures, the progress made to date, and the ongoing challenges.   The session will be held by ISD Expert Richard Kuchta, Senior Analyst in ISD Germany focusing on the research of extremism, information operations and platform regulation. His research looks at emerging information campaigns, platforms and technologies, combining digital research methods with a policy and regulatory perspective. 👉 Register now and join on Tuesday 21 July at 16:30 CEST:  https://lnkd.in/eMEewiXz ISD (Institute for Strategic Dialogue) ISD Germany

Co-Executive Director of ISD Germany Dr Julia Ebner's new book, The Language of Terrorists, is already turning heads. Building on a decade of research at ISD, Julia's book introduces a threat detection model known as the Violence Risk Index, exploring how language can help distinguish credible threats from violent extremist rhetoric—including emerging challenges posed by AI, blockchain and virtual reality.   In his review for The Literary Review, Lord David Anderson—former UK Independent Reviewer of Terrorism Legislation—describes Julia, and the book, as "a contemporary guide to this sinister and shifting world" of online extremism. "The Language of Terrorists showcases a clever multidisciplinary approach to the prediction of serious violence", Anderson writes. "More broadly, it is a reminder of how easily we can be given away by the words we choose to employ". The book builds on ISD's wider threat detection and prevention program, which seeks to understand the evolving landscape of online and offline harms in order to safeguard against the devastating fallout they can cause. Within its pages, Julia takes on one of the most urgent questions in that space: what can language tell us about the likelihood that rhetoric will translate into real-world violence? Read the full review is available here. https://lnkd.in/eSPnQ3vn  The Language of Terrorists will be published in August. https://lnkd.in/e-FWEVDQ   

ISD is excited to be involved in the establishment of NERON—the Network for European Research on Jewish Life and Antisemitism—led by the Institute for Jewish Policy Research (JPR) and European University Institute (EUI). Together, we're working to advance research on countering antisemitism & fostering Jewish life in Europe. Learn more about the initiative below ⬇️ https://lnkd.in/e_Cyx55u

The May 2026 attack on the Islamic Center of San Diego reflects a newer and increasingly visible form of neo-Nazi accelerationist violence: while extremist ideology remains central, the pathways into it are becoming increasingly diffuse and embedded in nihilistic online subcultures. New ISD research found that, although the attack was ideologically motivated and rooted in white supremacist accelerationism, the perpetrators' broader online footprints suggest that their pathways into violence were also shaped by other online environments, including mass shooter fandoms, misogynistic communities, online fandoms characterized by intense attachments to fictional characters and aesthetic-driven platforms where violent content can circulate beyond traditional extremist spheres. ISD's analysis also points to a fracture within the accelerationist landscape. While actors outside legacy accelerationist networks continue to draw on traditional accelerationist reference points, established networks often dismiss them as unserious, undisciplined or ideologically confused. This dynamic suggests that accelerationist violence is becoming more fragmented and less subject to control by legacy extremist communities. For policymakers and practitioners, these findings suggest that warning signs may increasingly emerge outside conventional extremist communities and across a wider range of online spaces. At the same time, the emergence of these newer pathways should not distract from the persistent threat posed by more traditional accelerationist networks, which remain active, organised and dangerous. 🔗 Read ISD's latest Dispatch. https://lnkd.in/e8GDE3wX

ISD’s CEO Sasha Havlicek features in the new documentary Miss Representation: Rise Up. The film, which premiered at the Tribeca Film Festival, confronts the dangers women and girls face online, including harassment, deepfakes, AI exploitation and algorithmic systems that have increasingly exploited the vulnerabilities of young women and men. The Representation Project Sasha will be joining a post-screening panel discussion with Jennifer Siebel Newsom this Sunday at the DC/DOX Film Festival in Washington. Tickets are available here: https://lnkd.in/eH2mhiFk #MissRepresentation #RiseUp #Tribeca2026 #GenderBasedViolence #OnlineSafety #TechFacilitatedGBV #ImageBasedAbuse

New data from ISD’s latest briefing on online domestic extremism in Canada highlights the narratives, platforms and events driving extremist activity online between December 2025 and February 2026. Drawing on analysis of 578,000+ posts from 437 active extremist accounts and channels across major social media platforms, the briefing examines key trends in hate speech, violent rhetoric and online mobilization. It also explores the topics that dominated extremist discussions, including immigration, gender identity, antisemitic conspiracy theories and geopolitical developments. This briefing is part of ISD’s ongoing monitoring of domestic extremism in Canada, providing evidence-based insights to support policymakers, practitioners and researchers working to strengthen social cohesion and resilience. Read the latest briefing here. https://lnkd.in/eyDEnc39