Framework for assurance of digital assets
The ADA project is maturing a framework and toolbox for the assurance of digital assets, taking a broad definition of the term as its point of departure. A digital asset may be composed of a combination of data, models, algorithms, software driven control systems, sensors & Internet of Things (IoT) devices, software applications & services, digital certificates, tokens and smart contracts. A common characteristic is that a digital asset is essential to an organisation’s business operations. There are two main categories: digitally-enabled physical systems and purely virtual assets. Digital assets unlock opportunities for new business models and improved business performance. However, the failure of a digital asset could result in substantial environmental damage, loss of revenue, loss of life, and ultimately a loss of trust.
Digital assets are often comprised of many components and elements, creating complex systems that require new approaches for hazard identification and safety assurance. Simulation is a key tool in identifying emergent properties such as safety and reliability.
The integration of Artificial Intelligence (AI), often as black-box systems with unknown input-output relationships, adds to the challenges of system complexity. DNV GL considers trust in AI as fundamental for its scalability and growth.
The aim of the ADA project is to establish a framework for assurance of digital assets, ultimately proving their trustworthiness. The project team has produced a set of white papers analysing the challenges of complex systems, and together with DNV GL Digital Solutions, will shortly publish a Recommended Practice for the Assurance of Advanced Algorithms. DNV GL has also outlined the characteristics that may make an AI system trustworthy and during 2020 will produce a proof of concept of the assurance framework and further explore how to leverage digital assets for sustainability.