Towards Modelling and Verification of Social Explainable AI