The observation of new physics events with large missing transverse energy at the LHC would potentially serve as evidence for the direct production of dark matter. A crucial step toward verifying such evidence is the measurement of the would-be dark matter mass. If, for example, the invisible particles are found to have masses consistent with zero, it may prove very challenging to ascertain whether light dark matter or neutrinos are being observed. We assume that new invisible particles are pair-produced in a -like topology and use two MT2-based methods to measure the masses of the particles associated with the missing energy. Instead of simulating events and backgrounds, we estimate the uncertainty associated with measuring the mass of the invisible particle by assuming a fixed value of the uncertainty associated with the location of the MT2 endpoint. We find that if this uncertainty associated with measuring the MT2 endpoints is, quite optimistically, O(1 GeV), the invisible particles must have masses greater than O(10 GeV) so they can be distinguished from massless ones at 95% CL. If the results from the CoGeNT, DAMA/LIBRA, and CRESST experiments have indeed revealed the existence of light dark matter with mass O(10 GeV), our results suggest that it may be difficult for the LHC to distinguish dark matter from neutrinos solely via mass measurements.