An argumentation-based approach to risk assessment

Reliable quantitative risk assessment requires the use of meaningful statistical data and well validated models for their interpretation. In certain domains, neither the data nor the validated models are available. Nevertheless, there may still be a requirement to generate a meaningful characterization of risk, presented in a form which reflects the reliability and accuracy of the data available. One such domain is the assessment of the potential carcinogenic risk in chemcial compounds, and there are many others. Work is described which is underway to provide a sound framework for the qualitative assessment of risk using a computer model of ‘argumentation’. This work is being applied to the development of computer-based support for the assessment of carcinogenic risk.