A framework for assessing AI ethics with applications to cybersecurity

In the last few years many scholars, public and private organizations have been involved in the definition of guidelines and frameworks for individuating the principles to adopt in the development and deployment of AI systems. Some authors, however, noted that the effectiveness of these guidelines or ethical codes on the developer’s community is very marginal. One of the obstacles that opposes to the effective implementation of ethical principles is the lack of an approach for solving tensions which arise when principles are applied. A possible solution to such an issue could be the adoption of a risk-based approach which is also advocated by many sources. To our knowledge, no concrete proposals have been presented in literature on how to perform a risk-based ethical assessment. In this paper we contribute to close this gap by introducing a framework based on a qualitative risk analysis approach for assessing the ethical impact underneath the introduction of an innovation either technological or organizational in a system. We will also show how the framework can be used for individuating suitable safeguards to adopt for balancing potential ethical infringements that the innovation may entail once implemented. Some case studies in the cybersecurity context are also described for showing the effectiveness of our approach.