Future Directions in Trusted and Self-Regulating Enterprise Intelligence Systems

Enterprise Intelligence Systems (EIS) use Artificial Intelligence (AI) techniques to automate decision-making at enterprise levels as a means of improving productivity, profitability, and competitiveness. Trust in such systems can lead to their wider adoption and use. Trust can motivate self-regulation, reducing dependence on external regulation or assurance and enabling faster, more efficient and less costly operation. This presentation discusses the foundations of trust and self-regulation in EIS, their architectural imperatives, practices for self-regulation, and suggestions for future research in the area of trust in enterprise intelligence systems. Questions addressed include:

- What are the conceptual underpinnings of trust in enterprise intelligence systems?

- What architecture is needed in enterprise intelligence systems to engender trust?

- What governs trust in enterprise intelligence systems in practice?

- How are trust and self-regulation of enterprise intelligence systems maintained during operation?

- What are the necessary conditions for and possible sources of trust and self-regulation in enterprise intelligence systems?

- How do these systems differ from other AI-based systems, such as algorithms used to predict the outcome of court cases, and what coping mechanisms are therefore needed?

- In what ways is the behaviour of enterprise intelligence systems distinct from AI-based systems for image classification, recommendation engines, and chatbots, and how are trust and self-regulation therefore supported?