A processor associated with a vehicle receives sensor data from a plurality of sensors associated with a vehicle, where each sensor generates data corresponding to a different parameter of a passenger in the vehicle. Based on the sensor data, one or more primitive emotional indications are generated. The processor applies a model to the one or more primitive emotional indications that when applied outputs a contextualized emotional indication associated with the passenger that includes an assessment of an emotional state of the passenger and a reason for the emotional state. A contextual response is selected based on the contextualized emotional indication, and the processor causes an output by the vehicle to enact the contextual response.