Disruption of naturally evolved spatial patterns of genetic variation and local adaptations is a growing concern in wildlife management and conservation. During the last decade, releases of native taxa with potentially non-native genotypes have received increased attention. This has mostly concerned conservation programs, but releases are also widely carried out to boost harvest opportunities. The mallard, Anas platyrhynchos, is one of few terrestrial migratory vertebrates subjected to large-scale releases for hunting purposes. It is the most numerous and widespread duck in the world, yet each year more than three million farmed mallard ducklings are released into the wild in the European Union alone to increase the harvestable population. This study aimed to determine the genetic effects of such large-scale releases of a native species, specifically if wild and released farmed mallards differ genetically among subpopulations in Europe, if there are signs of admixture between the two groups, if the genetic structure of the wild mallard population has changed since large-scale releases began in the 1970s, and if the current data matches global patterns across the Northern hemisphere. We used Bayesian clustering (Structure software) and Discriminant Analysis of Principal Components (DAPC) to analyze the genetic structure of historical and present-day wild (n = 171 and n = 209, respectively) as well as farmed (n = 211) mallards from six European countries as inferred by 360 single-nucleotide polymorphisms (SNPs). Both methods showed a clear genetic differentiation between wild and farmed mallards. Admixed individuals were found in the present-day wild population, implicating introgression of farmed genotypes into wild mallards despite low survival among released farmed mallards. Such cryptic introgression would alter the genetic composition of wild populations and may have unknown long-term consequences for conservation.