As part of the development of the advanced metering infrastructure (AMI), i.e., the "smart grid," power utilities are increasingly deploying residential meters that wirelessly communicate (either with devices in the home, with other meters, or with the utility). These meters may employ multiple antennas and radiate at different frequencies, ranging from 850 MHz to 2.4 GHz. Unlike radio-frequency (RF) exposure caused by cell phones, where the position of the phone relative to the body is somewhat fi xed, the position of a power meter relative to the body is rather unconstrained. In this work, we used the Finite-Difference Time-Domain (FDTD) Method to study the Specifi c Absorption Rate (SAR) produced in full anatomical models of humans when they were exposed to the RF fi elds produced by a wireless AMI meter, commonly referred to as a smart meter. Various scenarios were considered, most of which could be described as "beyond worst case." Despite this fact, only when an individual was extremely close to one particular radiating element and when one ignored the actual duty cycle of these meters did the SAR values exceed the published safety limits. When one accounts for the meter's true duty cycle or there was a realistic separation between the meter and an individual, all SAR values fell within safety limits.