LHC collisions can act as a source of photons in the initial state, in addition to the more common quark and gluon-initiated processes. Indeed, photon-intitiated production is a promising search channel for BSM states as well as probe of the EW couplings of the SM particles. Due to the colour singlet nature of the photon, a key feature of this process in proton-proton collisions is the possibility for leaving the protons intact and/or producing rapidity gaps in the final state. Indeed, the outgoing intact protons can be measured by dedicated `tagging' proton detectors in association with ATLAS and CMS. Moreover, the possibilities are not limited to proton collisions: in heavy ion collisions, the ions can act as a strong source of photon radiation, and the photon-initiated channel can play a significant role.
In this talk I will overview the current status and prospects for photon-initiated production at the LHC. I will discuss the theoretical foundations underlying the modelling of such processes and their implementation in a Monte-Carlo event generator. I will in particular demonstrate that the underlying theory is well understood, with limited sensitivity to unconstrained region of QCD due to the strong interaction of the colliding hadrons. We are therefore well justified in viewing such processes as photon-photon collisions, even if the devil is as always in the detail, as I will discuss.