What happens when you ask AI chatbots about a crisis or terror attack? While models are programmed to block violent content, real-world audits reveal a concerning phenomenon: “Reject then launder.”
Chatbots initially condemn the violence to pass safety filters, only to immediately validate and whitewash the underlying extremist ideology in their response. Crucially, this is not a real-time data gap where models simply lack confirmed facts during a breaking story. Through real-world case studies, including an audit of nine frontier LLMs on the 2011 Norway terror attack and social feed dynamics from the Bondi Beach shooting, this session proves the core issue is not a shortage of information, but a fundamental flaw in how LLMs process and reason through ideology.
In this webinar, speakers from Factiverse and SimPPL will dive into these real case study findings. Beyond LLM vulnerabilities, the session will explore how unverified viral posts and faulty AI outputs rapidly become the “accepted record” during high-stakes crises, providing practical, cost-effective frameworks for researchers, fact-checkers, and policymakers to audit these environments.
Speaker: Seán Jacob, Senior Project Manager at Factiverse
Seán Jacob is a Senior Project Manager at Factiverse, the Norwegian AI-powered claim detection company behind deployments at NRK and SVT. He is one of the leads on Factiverse’s LLM auditing projects, working with partner organisations to develop methodologies for testing how AI models handle conspiracy narratives and crisis-related misinformation.
Speaker: Dr. Swapneel Mehta, Cofounder of SimPPL
Atlantic Dialogues Emerging Leader, Google Research Innovator, and a Belfer Fellow. Swapneel previously helped build the BU/MIT Platform Governance Lab. Ph.D. at NYU’s Center for Data Science working on machine learning and causal inference with applications to social media and politics at CSMAP.
Moderator: Alexandre Alaphilippe, EU DisinfoLab
Alexandre Alaphilippe is the Executive Director and co-founder of the EU DisinfoLab. Since 2017, he has coordinated work on some of the organisation’s largest investigations into Information Operations linked to Russia, India and China. In 2022, he led the exposure of Doppelganger, which has been labeled as one of the largest information operation from Russia in the past years. He is a member of a number of working groups in Brussels and Washington DC linked to platform regulation, transatlantic relations, and hybrid threats, where he emphasises the role of civil society in maintaining democratic values. He has published papers for the Brookings Institution and his work has been featured on CNN, BBC, Le Monde and Politico.
This webinar is sponsored by Factiverse.
The opinions expressed are those of the speakers/authors and do not necessarily reflect the position of EU DisinfoLab. This webinar does not represent an endorsement by EU DisinfoLab of any organisation.