Environmental risk assessment often involves the use of models, the results of which are sometimes used in regulatory decisions or in drafting of legislation. Many mathematical models have been employed to quantify the chemical exposure that occurred for humans over a period. However, the high complexity and inherent heterogenity of chemical exposure is still a major challenge for the scientific community. Because of limited scientific data, interpretation of models often involves uncertainty. Modelling the uncertainty inbuilt in the exposure analysis can be solved using recent computational techniques like fuzzy logic and neural networks. In this paper, human health risk through inhalation and dermal exposure to benzene from vehicular emissions in the city of Christchurch, New Zealand, is assessed as an example of the application of a new hybrid approach to risk assessment. Major variables affecting the absorption of chemicals and key parameters for dispersion of chemical are considered in the model. The model has produced excellent results and showed that new artificial intelligence algorithms are a promising tool for risk assessment.