A main obstacle in the widespread application of anion exchange membranes (AEMs) and AEM fuel cells (AEMFCs), aside from high-pH stability issues, is the low conductivity of the membrane [1]. Despite advantages such as more facile oxygen reduction reaction kinetics [2], even high performing AEMFCs still exhibit power densities many times lower than their proton exchange counterparts [1-3]. Because AEM is inherently less conductive, it is important that their conductivity behavior is well understood in order to mitigate any factors that result in further decrease in the membrane’s conductivity, or at least design around them. Furthermore, potential applications for AEMs and AEMFCs involve exposing the AEM to carbon dioxide [2, 4, 5], which reacts to form carbonate and bicarbonate species and leads to a significant decrease in the membrane’s conductivity [1-7]. Any devices intending to operate an AEM in the carbonate state would benefit from a detailed understanding of these ions’ effect on membrane performance. Efforts to elucidate the conductivity behavior of AEMs have been mostly experimental, but some theoretical models have also been developed using various approximations [5, 10-13]. Some past models assume chemical equilibrium among ionic species, however in an operating AEM this may not be the case. Furthermore, most of the theoretical models describe the behavior of isolated AEMs, that is, in the absence of polarization and electrochemical reactions that supply and consume ions and influences the ion exchange process. This work will incorporate these effects in order to predict the conductivity of an operating AEM in the presence of CO2 and under various operational current densities and temperatures. The model itself is a direct extension of [7], in that it adapts the same reaction mechanism for CO2 conversion to carbonate/bicarbonate ions in the AEM. Additional fundamental insight into the self-purging mechanism and potential scenarios for the recovery of membrane conductivity will be discussed. Acknowledgements Financial support from the Army Research Office (award number W911NF-14-1-0298) is gratefully acknowledged.