How to build trust in an uncertain world: digital twins and beyond

How to build trust in an uncertain world: digital twins and beyond

In their recent seminal work “Radical Uncertainty”, former Bank of England Governor Lord (Mervyn) King and Oxford academic John Kay argue that despite the power of probability theory, simple percentages cannot predict what we really need to know about critical events like the Global Financial Crisis or (they were prescient) pandemics. They argue that a more sophisticated understanding of what leads to “radical uncertainty” is needed. We still need granularity, but we also need to grasp the big picture. The question we should ask, they say, is “what is going on here?”, in the real world, not just in the world of numerical probabilities

In other words, we have the numbers but we are missing the semantics: the narrative, meaning and context behind the raw data. If we don’t understand the meaning, we can’t trust the numbers to tell us what we need to know.

A related challenge faces those who are responsible for complex mission-critical systems on which today’s world increasingly depends. Whether in aerospace or complex financial services or border control, success or failure can be business-changing, even a matter of life and death. Such complex operations often depend on multiple parties committing to make a specific contribution to a “grand plan” which they may not fully understand or even have access to. It isn’t just the familiar challenge of getting legacy systems to talk to each other. It is also a question of whether the data in the system actually reflects an agreed interpretation of the real world, ie “what is going on here”. Even if the data is accurate, those who have to take key decisions are usually overloaded with more data than they can manage. Often, they won’t afterwards be able to see in detail what worked and what went wrong – something which is essential for compliance, liability and lesson learning. What’s happened is that we have the data but lose sight of the important story behind it. Even if we understand the meaning of individual pieces of data – the semantics – we don’t understand “what’s going on here” when all the data interacts – the pragmatics.

An intriguing way to approach the challenge of semantics and pragmatics is by thinking about how the physical and logical worlds are connected, which remains one of the big challenges of digitisation. The idea isn’t new. The term “digital twin” was introduced by John Vickers of NASA in a 2010 Roadmap Report coincidentally around the same time as the introduction of the bitcoin blockchain, though the concept dates back to 2002 and Michael Grieves work on product lifecycle management. A Digital Twin provides a digital replica of actual physical assets, processes, people, and systems that can be manipulated, simulated and modified in the digital world without expensive real-world interventions. Any given state in the digital world has a correspondence in the real physical world.

But we can now be more ambitious about what we can do by linking the real and digital worlds. What we’re calling a “Digital Cognate” provides a digital entity; an asset uniquely linked to an object of real value that reflects the evolution, custodianship and transfer of value that is immutably locked in an historical record that is private and secure to tampering.

In a world of Radical Uncertainty digital cognates provide a mechanism for establishing trust. Attaching a physical marker with a unique identifier to each object in a value chain and linking it immutably to a digital cognate with its own unique identifier embeds trust in the veracity of each data point as it enters the system. Data can be extracted and manipulated according to the specific objectives of the operation. In this way, we can identify the information which is crucial to success or risk and have confidence that the outcome reflects reality, however long or complex the value chain.

An obvious use case for this novel application of digital cognates is complex and safety-critical manufacturing, such as aerospace. Each component in a manufacturing process is tagged with a physical identifier, eg RFID or NFC chip. The unique identifier links the object to the digital twin and the data is hashed. Other parameters include operator access and permissions. The user then defines which information is critical to output maximisation and risk management. This data is extracted to provide both real-time dashboards and an immutable record for regulatory and other purposes.

“Digital cognate” suggests that there is usually a linked physical object. But the digital twin – UID approach can also be applied to “desk-based” operations such as in financial services or the attribution of digital assets of value. Take, for example, insurance product development and model validation. Each product is given a UID which is also attached to the matching digital cognate, enabling the same trust in the integrity of the process.

With any innovative technology, the first question we ask is “so what?” Is there a ROI for me? In this case, there are two answers – one simple “yes” and a potentially more exciting “yes”.

The simple “yes” is that confidence in the veracity of the input data in a complex mission-critical system enables performance optimisation, risk management as well as cost reduction in processing. When you really know “what’s going on here”, you can improve quality, performance and customer satisfaction, reduce downtime and save on overheads. And when you know “what went on here”, you can meet compliance obligations more efficiently, simplify audit and protect reputation when things go wrong. Trust is assured by design through the Digital Cognate.

The more exciting “yes” comes from the even greater potential of having – for the first time -data we can really trust. The Digital Cognate ensures trust in a multi-party ecosystem, offering the potential for performance optimisation and risk management. But we can go further. The next stage is to extend the trust to the semantics and pragmatics, taking control of the context and providing confidence in the integrity of complex processes even to third parties. This is when we can really speak of smart legal contracts.

To embed meaning into the digital cognate calls for language that is both human- and machine- readable, so that we can incorporate not just data but the intent behind it. Our vision is to use our version of Ricardian contracts as a mechanism to relate meaning and intent to data. Ricardian contracts define a legally valid and digitally connected “document” (in a general sense) to a physical object or value in a form that can be executed by machine. In this way, the intent of the data is integrated into the digital cognate which is attached to the evolution in the digital domain. When it is transposed back into the physical domain, the Ricardian contract reliably confirms the meaning to the recipient.

Increasingly “smart” comes to mean “digital”. But we can only really trust data if we can be sure where it came from in the physical, human, world. Now digital cognates can give us “smart” we can really trust.

David Lowe, Chief Technology Officer

David Landsman, Non-Executive Director