We present predictions for the evolution of radio emission from Active Galactic Nuclei (AGNs). We use a model that follows the evolution of Supermassive Black Hole (SMBH) masses and spins, within the latest version of the GALFORM semi-analytic model of galaxy formation. We use a Blandford-Znajek type model to calculate the power of the relativistic jets produced by black hole accretion discs, and a scaling model to calculate radio luminosities. First, we present the predicted evolution of the jet power distribution, finding that this is dominated by objects fuelled by hot halo accretion and an ADAF accretion state for jet powers above $10^{32}mathrm{W}$ at $z=0$, with the contribution from objects fuelled by starbursts and in a thin disc accretion state being more important for lower jet powers at $z=0$ and at all jet powers at high redshifts ($zgeq3$). We then present the evolution of the jet power density from the model. The model is consistent with current observational estimates of jet powers from radio luminosities, once we allow for the significant uncertainties in these observational estimates. Next, we calibrate the model for radio emission to a range of observational estimates of the $z=0$ radio luminosity function. We compare the evolution of the model radio luminosity function to observational estimates for $0<z<6$, finding that the predicted evolution is similar to that observed. Finally, we explore recalibrating the model to reproduce luminosity functions of core radio emission, finding that the model is in approximate agreement with the observations.