Framework for Managing Risk in Industrial AI Systems
Lars Meloe, Principal Consultant, DNV
In this presentation, Lars Meloe, Principal Consultant at DNV, introduces a practical framework for developing and operating industrial AI systems safely and reliably.
The framework begins by balancing potential business value, including increased production, greater efficiency and faster decisions, against development, data, organisational change and maintenance costs. It follows AI from the initial idea through procurement, development, technology qualification and operation.
Lars explains that AI risks can emerge at several interconnected layers: the AI component, the wider AI-enabled system, its operational context and organisational governance. Trust therefore depends on more than model performance alone.
The presentation shows how risk management and assurance requirements can be applied across each layer. Proven approaches such as assurance cases and barrier management can help organisations define acceptance criteria, demonstrate control and build confidence in high-risk industrial applications.
Explore the full presentation below to understand why AI assurance must become a continuous, adaptive and evidence-based process that evolves alongside the technology, operating environment and associated risks.
