Description: AI systems are being rapidly deployed across industries to drive economic growth but they remain vulnerable to exploitation by a wide range of malicious actors, including Terrorist and Extremist Entities, Hostile Foreign States, and Organized Crime. While these misuse cases remain a critical challenge for both frontier models and downstream AI applications, machine learning tools can also help us detect and prevent malicious actors from producing high-severity harms at scale. Through real-world observed case studies, these misuse methods and their implications for AI security and international security will be examined in detail. Following this discussion, new methods and technological solutions being used to prevent malicious actors from exploiting AI systems will be examined with best practices highlighted for researchers, civil society, and industry. Despite recent shifts in AI Safety, preventing high-severity harms, including terrorist and violent extremist content (TVEC), chemical, biological, radiological, and nuclear (CBRN) risks, foreign influence operations, ransomware, targeted phishing campaigns, and scams remain top priorities for governments and industry alike. While new technological solutions offer new opportunities to counter these threats at scale, a broad multi-stakeholder approach remains critical to ensuring the economic benefits of AI can be harnessed while mitigating security risks.
Bio: Broderick McDonald is an Associate Fellow at Kings College London’s International Center for the Study of Radicalisation working at the intersection of conflict, extremism, and technology. His research and commentary has appeared in The New York Times, The Wall Street Journal, Foreign Affairs, Financial Times, The Guardian, BBC, Al Jazeera, France24, L’Orient-Le Jour, Middle East Eye, The Economic Times, The National News, Prospect Magazine, Al Arabiya, CS Monitor, and The Globe & Mail and has provided expert analysis for a range of international news broadcasters including ABC News, BBC World News, BBC America, CBC News, Good Morning America, France24, and Al Jazeera News. Prior to this, he served as an Advisor to the Government of Canada and was a Fellow with the United Nations Alliance of Civilizations (UNAOC). He previously lived in the Middle East and has conducted extensive quantitative and qualitative fieldwork with armed combatants and foreign terrorist fighters from ISIS, HTS, and other armed non-state actors. He currently serves on the Global Internet Forum to Counter Terrorism (GIFCT)‘s Independent Advisory Committee, the Aspen Institute UK’s RLF Advisory Board, and the GLOCA Board of Advisors. Alongside his research, Broderick has advised policymakers, parliamentarians, intelligence agencies, international prosecutors, NGOs, social media platforms, and AI developers on emerging technologies and security threats from terrorist & extremist entities, organized crime, and hostile foreign states.
