Abstract: Previous studies have used climate model simulations with perturbed sea ice covers to assess the climate impact of future Arctic sea ice loss. The results of these studies suggest that Arctic sea ice loss will cause substantial warming both in the Artic and beyond. The approaches used in these simulations can be broadly categorized into three methodologies: modifying the surface albedo, adding a ghost flux to the sea ice module, and nudging. Here we show that all three methodologies effectively add heat to the Arctic in order to melt the sea ice. This causes an artificial warming signal that is added to the warming that occurs due to sea ice loss alone. We illustrate this using an idealized climate model. In this model, the annual-mean warming due to sea ice loss alone can be directly calculated. We compare this with the warming that would be attributed to sea ice loss using each of the three methodologies in the idealized model. The results show that each methodology substantially overestimates the warming due to sea ice loss alone, overestimating the surface warming throughout the Northern Hemisphere by a factor of 1.5-2. Hence these results suggest that previous studies have overestimated the impact of sea ice loss on the climate system.