Case Studies on AI Model Manipulation for FIMI Conference Brussels

I'll be popping over to Brussels next week to talk at the big 2026 European External Action Service conference on foreign information manipulation from Monday to Wednesday. I know the event can be quite helter skelter, but if you're around, ping me! Let's grab a coffee or (Belgian) beer. I'm going to be talking about the future of FIMI in the AI age - and one of my big focuses this year is on the manipulation *of* AI models in covert and scaled ways in order to have influence. We have really focussed too much on deepfakes so far, but my sense is that the race to influence and mould prompt responses at scale is the real gold rush of influence, and that the scramble to win it is really is going to be one of the defining characteristics of FIMI from hereon in. I am very keen to hear from anyone that might point me towards case studies. It doesn't have to be geopolitics - it might be in the area of market manipulation, or brands and reputations. And it doesn't have to be proven or certain; I'm as keen to investigate examples which are emergent, unproven, and ambiguous. Thank you!