A novel technique is being developed to estimate the effective dose of a neutron field based on the distribution of neutron captures in a scintillator. Using Monte Carlo techniques, a number of monoenergetic neutron source energies and locations were modelled and their neutron capture response was recorded. Using back propagation Artificial Neural Networks (ANN) the energy and incident direction of the neutron field was predicted from the distribution of neutron captures within a 6Li-loaded liquid scintillator. Using this proposed technique, the effective dose of 252Cf, 241AmBe and 241AmLi neutron fields was estimated to within 30% for four perpendicular angles in the horizontal plane. Initial theoretical investigations show that this technique holds some promise for real-time estimation of the effective dose of a neutron field.