The Kleinmann--Low nebula in Orion, the closest region of massive star formation, harbors Source I, whose nature is under debate. Knowledge of this source may have profound implications for our understanding of the energetics of the hot core in Orion KL since it might be the main heating source in the region. The spectral energy distribution of this source in the radio is characterized by a positive spectral index close to 2, which is consistent with (i) thermal bremsstrahlung emission of ionized hydrogen gas produced by a central massive protostar, or (ii) photospheric bremsstrahlung emission produced by electrons when deflected by the interaction with neutral and molecular hydrogen like Mira-like variable stars. If ionized hydrogen gas were responsible for the observed continuum emission, its modeling would predict detectable emission from hydrogen radio recombination lines (RRLs). However, our SMA observations were obtained with a high enough sensitivity to rule out that the radio continuum emission arises from a dense hypercompact H II region because the H26alpha line would have been detected, in contrast with our observations. To explain the observational constraints, we investigate further the nature of the radio continuum emission from source I. We have compared available radio continuum data with the predictions from our upgraded non-LTE 3D radiative transfer model, MOdel for REcombination LInes, to show that radio continuum fluxes and sizes can only be reproduced by assuming both dust and bremsstrahlung emission from neutral gas. The dust emission contribution is significant at nu >= 43 GHz. In addition, our RRL peak intensity predictions for the ionized metals case are consistent with the nondetection of Na and K RRLs at millimeter and submillimeter wavelengths.
