A half-day workshop that treats mis/disinformation as an end-to-end socio-technical risk, from data, training and alignment to deployment, dissemination, detection and societal impact.
Disinformation and misinformation are long-standing societal challenges. The rapid development and wide availability of generative AI systems have changed their scale, speed and impact: highly convincing multimodal content can now be produced at low cost, while countermeasure technologies struggle to keep pace.
Existing venues such as SafeAI, De-Factify, MAD, ROMCIR and COMPASS each emphasize one aspect of the problem: general AI safety, multimedia forensics, fact-checking, information retrieval or platform moderation. This workshop connects these strands through the lens of generative AI safety, bringing together technical work on interpretability, multimodal detection, provenance, watermarking, red-teaming and agentic evaluation with research on user behavior, cultural bias, vulnerability and collective harm.
As an inaugural event, the goal is to move beyond the detection of isolated false claims or manipulated artifacts, toward understanding and mitigating the broader systems through which generative AI can influence what people believe, share and trust.
We welcome contributions drawing on both computer science and the social sciences. Topics include, but are not limited to:
Submission site, formatting template and publication details will be announced here.
Half-day workshop. The date (February 22 or 23, 2027) and the final schedule will be confirmed by AAAI.
The committee spans decentralised agentic systems and data governance, explainable AI and mechanistic interpretability, information integrity and responsible NLP, and biomedical and health text mining, bridging industry and academia across Europe and North America.
Senior Researcher in the Security Science Laboratory of Fujitsu Research of Europe, working within the Frontria Consortium on decentralised agentic systems and data governance. PhD from Durham University. Co-organised the Machine Learning for Multiscale Processes Workshop at ICLR 2025, the AI4X 2025 conference and the AI for Science and Nobel Turing Challenge Conference 2024. eleonore.gaudin@fujitsu.com
Team Leader of the Model Explainability and Alignment team at Intesa Sanpaolo AI Research, where he also leads an industrial project on the safety of the bank's LLMs. PhD in AI (Turin & City University London). Chairs the XAI Session of the International Neural Networks Society; lead organizer of the Interpretability & Alignment of Financial Models workshop at ICAIF '26. alan.perotti@intesasanpaolo.com
Postdoctoral researcher at Mila working on information integrity and AI for social good. Leads the Veracity project, highlighted in the Canadian AI Strategy report. PhD from Université de Montréal. Organized the AiOfAi workshops at IJCAI 2023 and The ACM Web Conference 2026. dorsaf.sallami@umontreal.ca
Professor of Computer Science at the University of Manchester and Director of the UK National Centre for Text Mining; ELLIS fellow. Works on NLP for information extraction, summarisation, emotion and misinformation detection. Has organised BioNLP and CL4Health at ACL, EMNLP and COLING; co-founder of SIGBioMed and SIG-FinTech (ACL). sophia.ananiadou@manchester.ac.uk