The doubly-magic nucleus $^{16}$O has a small neutron capture cross section of just a few tens of microbarn in the astrophysical energy region. Despite of this, $^{16}$O plays an important role as neutron poison in the astrophysical slow neutron capture ($s$) process due to its high abundance. We present in this paper a re-evaluation of the available experimental data for $^{16}$O($n,\gamma$)$^{17}$O and derive a new recommendation for the Maxwellian-averaged cross sections (MACS) between $kT$= 5$-$100 keV. Our new recommendations are lower up to $kT$= 60 keV compared to the previously recommended values but up to 14\% higher at $kT$= 100 keV. We explore the impact of this different energy dependence on the weak $s$-process during core helium- ($kT$= 26 keV) and shell carbon burning ($kT$= 90 keV) in massive stars where $^{16}$O is the most abundant isotope.