The recent detection of a neutron star merger by the LIGO collaboration has renewed interest in laboratory studies of r-process elements. Accurate modeling and interpretation of the electromagnetic transients following a neutron star merger requires computationally expensive calculations of both the structure and opacity of all trans-iron elements present in the ejecta. To date, the necessary atomic data to benchmark the structure codes used to provide this atomic data is incomplete or in some cases absent entirely. Within the available laboratory studies of gold and its ion, there are conflicting reports on the emission lines and level structures. We present a new study of Au I and Au II lines and levels by exposing a solid gold target to plasma in the Compact Toroidal Hybrid (CTH) experiment at Auburn University. In Au I, 86 lines are observed, 45 which are unreported in the literature, and the energies of 18 6s6p levels and 16 of the 18 known 6s6d levels are corroborated by a least-squares level energy optimization. In Au II, 76 emission lines are observed, and 51 of the lines are unreported in the literature. For both Au I and II, the new lines predominantly originate from the most energetic of the known levels, and over half of the new Au II lines have wavelengths longer than 300 nm. For the estimated electron parameters of CTH plasmas at the gold target (ne ~ 1012 cm-3, Te ~ 10 eV), two-electron transitions in Au I and II are similar in intensity to LS-allowed one-electron transitions. These results are expected to aid in the analysis of both neutron star merger and chemically peculiar stellar spectra.