In this paper a framework is proposed for the adaptive control of robotic manipulators which combines parametric adaptive control with Artificial Neural Network (ANN)-based compensation of dynamic uncertainties like friction. The proposed method utilizes a passivity-based parametric adaptive control approach and makes use of the ANN models as generic identifiers to compensate for unmodelled friction effects. Unlike many approaches for ANN based control in the literature, parameter update equations for the ANN model and for the parametric adaptive model are driven by both the tracking error and the system identification error. A stability analysis is given based on the passivity properties of the manipulator dynamics. The methodology is successfully tested for the control of a Direct Drive SCARA arm and performance is compared with standard adaptive control schemes.