1: \begin{abstract}

2: Stars are known to be more active when they are young, resulting in a strong correlation between age and photometric variability. The amplitude variation between stars of a given age is large, but the age-variability relation becomes strong over large groups of stars. We explore this relation using the excess photometric uncertainty in \gaia\ photometry (\varg, \varb, and \varr) as a proxy for variability. The metrics follow a Skumanich-like relation, scaling as $\simeq t^{-0.4}$. By calibrating against a set of associations with known ages, we show how \var\ of population members can predict group ages within 10-20\% for associations younger than $\simeq$2.5 Gyr. In practice, age uncertainties are larger, primarily due to finite group size. The index is most useful at the youngest ages ($<$100\,Myr), where the uncertainties are comparable to or better than those derived from a color-magnitude diagram. The index is also widely available, easy to calculate, and can be used at intermediate ages where there are few or no pre- or post-main-sequence stars. We further show how \var\ can be used to find new associations and test if a group of co-moving stars is a real co-eval population. We apply our methods on the Theia groups within 350\,pc and find $\gtrsim$90\% are inconsistent with drawing stars from the field and $\simeq$80\% have variability ages consistent with those derived from the CMD. Our findings suggest the great majority of these groups contain real populations.

3: \end{abstract}

4: