The unified model for Active Galactic Nuclei (AGN) accounts for a variety of observational differences in terms of viewing geometry alone. However, from the fitting of high spatial resolution infrared (IR) data with clumpy torus models, it has been hinted that the immediate dusty surroundings of Type-1 and 2 Seyfert nuclei might be intrinsically different in terms of covering factor (torus width and number of clouds). Moreover, these torus covering factors also showed variations among objects belonging to the same type, in contradiction with simple unification. Interestingly, these intrinsic differences in Seyfert tori could explain, for example, the lack of broad optical lines in the polarized spectra of about half of the brightest Seyfert 2 galaxies. On the other hand, recent IR interferometry studies have revealed that, in at least four Seyfert galaxies, the mid-IR emission is elongated in the polar direction. These results are difficult to reconcile with unified models, which claim that the bulk of the mid-IR emission comes from the torus. In this invited contribution I summarize the latest results on high angular resolution IR studies of AGN, which constitute a crucial test for AGN unification. These results include those from the mid-infrared instrument CanariCam on the 10.4 m Gran Telescopio CANARIAS (GTC), which are starting to be published by the CanariCam AGN team, Los Piratas (https://sites.google.com/site/piratasrelatedpublications).