We reinvestigate the putative giant spin splitting at the surface of SrTiO$_3$ reported by Santander-Syro $et~al.$ [Nature Mat. 13, 1085 (2014)]. Our spin- and angle-resolved photoemission experiments on (001) oriented surfaces supporting a two-dimensional electron liquid with high carrier density show no detectable spin polarization in the photocurrent. We demonstrate that this result excludes a giant spin splitting while it is fully consistent with the unconventional Rashba-like splitting seen in band structure calculations that reproduce the experimentally observed ladder of quantum confined subbands.