Intelligent virtual assistants (IVAs): trust and zero trust

Intelligent virtual assistants (IVAs) for cyber security appear to offer promising solutions to tackle the problem of gaps in the future cyber security workforce. However, this paper argues that a problem emerges as artificial intelligence (AI) partners take on their roles. In AI - implicit trust is the norm, yet in cyber security, zero trust protocols are now mandated. The contribution of this conceptual paper is firstly to present an argument for the deployment of zero trust protocols to effectively manage our future AI partners, and secondly, to set out the first steps in a process to assess the operationalization of zero trust. By leveraging well-established theory on trust from organization and conflict management studies, zero trust can be evaluated. The zero trust assessment involves determining: propensity to trust; experience of trust; a trust assessment based on ability, benevolence, and integrity; followed by a study of; acceptance of vulnerability and of risk taking. Implications for practitioners, policymakers and academics include an argument that the deployment, assessment, and management of a zero trust posture will promote explainable, trustworthy, and secure AI. Further studies are called for.