Impressive limits of detection have been achieved in the literature. However, some of these methods have a poor sensitivity and the size of the voltammetric peaks generated at low arsenic concentrations are often too close to the background forsatisfactoryextrapolation.WhilsttheLODwereportherein is well within the range defined by the WHO regulations, the principal aim of this report is to show that, by supporting the GNPs on a random array of carbon nanotubes in this manner, the sensitivity towards arsenic is greatly improved to a value almost 20 times higher than any electrochemical method reported to date. We envisage that incorporating the AuCNTs into existing arsenic detection technology, where problems from interferences have already been reduced or overcome wouldsignificantlyaddtothesensitivityandreliabilityofsuch sensors, with clearly observable analytical signals produced even at trace levels of arsenic. Furthermore, it was observed that the AuCNTs produced reproducible results even after they had been stored for more then 10 months in ambient conditions which suggests this material and procedure is possibly ideal forin-fieldar senic detection where long storage time and immediate deployment are also desirable.