This work explores personalities of individuals at risk of Islamist radicalisation. Previous studies have successfully identified radicalised individuals on social media. However, a more prophylactic approach to counter radical communities would be to detect individuals prior to getting radicalised. A manually annotated dataset of 15,195 tweets from 259 persons believed to be at risk of radicalisation was classified according to Big 5 personality traits and compared to the same number of non-radical users. The results show that risk users can successfully be identified and exhibit distributional differences compared to non-radical users for Extraversion and Openness.