Prey face a conflict between acquiring energy and avoiding predators and use both direct and indirect cues to assess predation risk. Illumination, an indirect cue, influences nocturnal rodent foraging behaviour. New Zealand holds no native rodent species but has introduced mice (Mus musculus) that severely impair native biodiversity. We used Giving-Up Densities (GUDs) and observations of foraging frequency and duration to assess if artificial light induces risk avoidance behaviour in mice and could limit their activity. We found both captive (wild strain) mice in outdoor pens and wild mice within a pest fenced sanctuary (Maungatautari, New Zealand) displayed avoidance behaviour in response to illumination. In captivity, total foraging effort was similar across lit and unlit pens but mice displayed a strong preference for removing seeds from dark control areas (mean: 15.33 SD: +/-11.64 per 3.5 hours) over illuminated areas (2.00 +/-3.44). Wild mice also removed fewer seeds from illuminated areas (0.42 +/-1.00 per 12 hours) compared to controls (6.67 +/-9.20). Captive mice spent less than 1.0% of available time at illuminated areas, versus 11.3% at controls; visited the lit areas less than control areas (12.00 +/- 9.77 versus 29.00 +/-21.58 visits respectively); and spent less time per visit at illuminated versus control areas (8.17 +/-7.83 versus 44.83 +/-87.52 seconds per visit respectively). Illumination could provide protection at ecologically sensitive sites, damaged exclusion fences awaiting repair, fence terminus zones of peninsula sanctuaries and shipping docks that service offshore islands. We promote the hypothesis that the tendency of mice to avoid illumination could be a useful conservation tool, and advance knowledge of risk assessment and foraging under perceived danger.