Artificial light at night (ALAN) is a pervasive environmental pollutant that disrupts circadian rhythms, behaviour, and ecological interactions across taxa. Bats are particularly vulnerable, as illumination increases perceived predation risk and interferes with nocturnal foraging. While the spatial and spectral effects of ALAN have been documented, the extent to which prior exposure to ALAN modulates behavioural responses remains unclear. We experimentally tested whether bat responses to a novel lighting event differ between chronically illuminated and naturally dark environments. Using a paired-night illumination design, we compared bat activity and species richness under dark and lit conditions at 72 spatially independent sites in two urban parks chronically exposed to ALAN and in two structurally similar, naturally dark woodlands in Italy and Montenegro. Across 214,825 bat passes representing nine taxa/species groups, artificial light reduced the activity of light-sensitive species (Rhinolophus ferrumequinum, R. hipposideros, Myotis spp.) far more strongly in dark woodlands than in urban parks, indicating context-dependent attenuation of light responses in urban environments. Other taxa were either consistently depressed by light (Plecotus spp., Nyctaloids) or unaffected by it (pipistrelles, Tadarida teniotis). Species richness declined under light regardless of context. Overall, our results show that bat responses to ALAN depend on landscape context, particularly for light-averse taxa. This highlights that behavioural attenuation in illuminated environments may reflect ecological filtering, behavioural plasticity, or their combination, with implications for managing light pollution and conserving nocturnal biodiversity.