Modelling the efficacy of auto-internet warnings to reduce demand for child exploitation materials

A number of proposals have been made over the years to implement notification systems to modify user behaviour, by sensitising users to the fact that their activities are not anonymous, and that further consequences may follow from future detections of illicit activity. While these systems can be automated to a large extent, there is a degree of manual processing required, so the cost-effectiveness and potential user coverage of such controls is critical. In this paper, we consider the problem of sensitising entrenched paedophiles who search for and download large amounts of Child Exploitation Material (CEM). Some countries, like New Zealand, operate a centralised censorship system which could be used to issue notifications when entrenched paedophiles search for CEM. We develop a statistical model to determine how many notices would need to be sent to entrenched paedophiles to ensure that they receive at least one notification over a 12 month period. The estimate of CEM viewers is based on actual data from the New Zealand internet filter. The modelling results indicate that sending 9,880 notices would result in entrenched paedophiles receiving at least one notice; for average CEM users, 53.27% of users would receive at least one notice within 12 months.