Observations of quasar absorption spectra have suggested a variation of the fine structure constant α. This variation is a generic feature of many unified models of particle physics beyond the standard model, and hence, is a profound result. However, the most statistically significant portion of this sample involves the ratio of Mg to Fe wavelength shifts using the many-multiplet (MM) method. Recent calculations of low-metallicity intermediate-mass AGB stars are particularly important in this regard because they are shown to be copious producers of the heavy Mg isotopes. We have implemented these yields into a chemical evolution model and show that the ensuing isotope distribution of Mg can account for the observed α variation. These observations of quasar absorption spectra can be used to probe the nucleosynthetic history of high redshift (z > 0.5), low metallicity systems. This analysis and other abundances of low metallicity systems strengthen the mounting evidence that early star formation may have been influenced by an enhanced population of intermediate-mass stars. These intermediate-mass stars also affect other elemental and isotopic abundances, particularly CNO cycle elements. We compare our model to independent measurements of the C N, and O elements in Lyman-alpha clouds and show that the scenario we propose is consistent with observations.