Aerial surveys are a recognised technique to identify the presence and abundance of marine animals. However, the capability of aerial observers to reliably sight coastal sharks has not been previously assessed, nor have differences in sighting rates between aircraft types been examined. In this study we investigated the ability of observers in fixed-wing and helicopter aircraft to sight 2.5 m artificial shark analogues placed at known depths and positions. Initial tests revealed that the shark analogues could only be detected at shallow depths, averaging only 2.5 m and 2.7 m below the water surface for observers in fixed-wing and helicopter aircraft, respectively. We then deployed analogues at shallower depths along a 5 km-long grid, and assessed their sightability to aircraft observers through a series of transects flown within 500 m. Analogues were seen infrequently from all distances, with overall sighting rates of only 12.5% and 17.1% for fixed-wing and helicopter observers, respectively. Although helicopter observers had consistently higher success rates of sighting analogues within 250 m of their flight path, neither aircraft observers sighted more than 9% of analogues deployed over 300 m from their flight paths. Modelling of sighting rates against environmental and experimental variables indicated that observations were affected by distance, aircraft type, sun glare and sea conditions, while the range of water turbidities observed had no effect. We conclude that aerial observers have limited ability to detect the presence of submerged animals such as sharks, particularly when the sharks are deeper than ∼2.6 m, or over 300 m distant from the aircraft's flight path, especially during sunny or windy days. The low rates of detections found in this study cast serious doubts on the use of aerial beach patrols as an effective early-warning system to prevent shark attacks.