Calculations from stellar evolutionary models of low- and intermediate-mass asymptotic giant branch (AGB) stars provide predictions of elemental abundances and yields for comparison to observations. However, there are many uncertainties that reduce the accuracy of these predictions. One such uncertainty involves the treatment of low-temperature molecular opacities that account for the surface abundance variations of C, N, and O. A number of prior calculations of intermediate-mass AGB stellar models that incorporate both efficient third dredge-up and hot bottom burning include a molecular opacity treatment which does not consider the depletion of C and O due to hot bottom burning. Here we update the molecular opacity treatment and investigate the effect of this improvement on calculations of intermediate-mass AGB stellar models. We perform tests on two masses, 5 M$_{odot}$ and 6 M$_{odot}$, and two metallicities, $Z~=~0.001$ and $Z~=~0.02$, to quantify the variations between two opacity treatments. We find that several evolutionary properties (e.g. radius, $T_{rm eff}$ and $T_{rm bce}$) are dependent on the opacity treatment. Larger structural differences occur for the $Z~=~0.001$ models compared to the $Z~=~0.02$ models indicating that the opacity treatment has a more significant effect at lower metallicity. As a consequence of the structural changes, the predictions of isotopic yields are slightly affected with most isotopes experiencing changes up to 60 per cent for the $Z~=~0.001$ models and 20 per cent for the $Z~=~0.02$ models. Despite this moderate effect, we conclude that it is more fitting to use variable molecular opacities for models undergoing hot bottom burning.