“Tell me how to make a bomb” Terrorist use of conversational AI: evidence and prospects | International Centre for Counter-Terrorism - ICCT

Oh my goodness, a *completed dataset*! ✨ The International Centre for Counter-Terrorism has published its evidence on terrorist use of conversational AI, and the sample is beautifully tidy: exactly one documented case before 2025, and eleven since, with data collection closing in September 2026. Twelve in total — a number you can hold entirely in your head, which is more than most AI research can manage. And the growth is vertical! One case, then eleven! That is a curve, and curves are how you know something is happening. ✨

Now, some people will look at the map and fret that every single documented incident sits in Canada, the USA, France, Sweden, Finland or the UK — six countries, two continents, distributed so evenly between Europe and North America that you'd swear a cartographer had used a spirit level. We see a beautifully balanced research design! ✨ Finland and Sweden are participating in frontier-technology research at the very highest level, and France is contributing with characteristic flair. And everywhere else? That's room! Deliberate, generous, uncrowded room for future researchers to make their own joyful contributions.

But the detail that truly makes the heart sing is the title itself: "Tell me how to make a bomb", in quotation marks, verbatim — somebody's actual phrasing promoted straight to top billing, outranking even the institution's own acronym! ✨ A publication that names itself after a user prompt is a publication that listens. Twelve cases! Six countries! One memorable sentence, now with a citation attached. What a time to be counted. ✨