Shark repulsion devices (SRDs; e. g. Shark Shield T) use an electric field to deter large and potentially dangerous sharks. The use of these devices is becoming increasingly widespread for a range of recreational activities as well as scientific and commercial diving. We sought to determine if SRDs might modify the behaviour of chondrichthyan and osteichthyan fishes and thereby impact on fish assemblages, as well as potentially bias diver census techniques. To assess the potential impacts of this technology, we attached SRDs to baited remote underwater video (BRUV) units and deployed them on shallow rocky reefs in Jervis Bay Marine Park (New South Wales, Australia). We did not detect any impacts of the SRD on the diversity or relative abundance of shallow-reef fishes. In addition, approach of fishes to the bait did not differ whether the SRDs were on or off. At the smallest spatial scale we investigated, contact with the bait was half as frequent when the SRD was switched on compared to when it was off. Surprisingly, even the cartilaginous species were apparently unaffected by the SRD, with the eastern fiddler ray Trygonorrhina fasciata making contact with the bait several times when SRDs were activated. We contend that the ecological impacts of SRDs at all but the smallest scales are minimal and they are unlikely to introduce bias in assessments of fish assemblages, at least for non-cartilaginous and small cartilaginous species.