Blockchains, digital credentials and trust

Or, the need to talk about architecture

The number of conversations, Slack messages and media pieces around digital credentials continue to swell. This means that there is a growing audience who are learning about ‘digital credentials’, ‘verifiability’ or even ‘open standards’. And this also means that precise, accurate and specific messaging which can respond to any FUD around the topic is the need of the hour.

The confidence in the adoption of digital credentials will rise only when more success stories are widely available. These stories of success would need good implementation playbooks. In turn these need robust foundational patterns - of governance around the layers of the stack.

The draft document focuses on the need to enable a national level blockchain initiative and includes topics around APIs which enable citizen services to come up. However, what it does not quite get into is about data semantics - the entire cycle of data input, data processing and data storage. Along with this, there is the absence of raising the awareness around designing systems which take care of data types to write to an immutable store.

The model of business enabled by blockchains depend on seamless data exchange. Thus there is an urgent need to focus on the data layer including elements of how to mask data items which are designated to be personal identifiers. This is a cautionary tale around how data systems can become colossal failures.

  • The Ada Lovelace Institute has a wonderful conversation around ‘The history and uses of vaccine passports and COVID status apps’

As digitally verifiable, secure and portable vaccination credentials get more mainstream attention there is a need to understand this in the historical context of the document. There are legitimate concerns around poorly designed digital credentials and how they can cause devastating issues around agency, fairness and equity. And when a number of proposals come up combining biometrics (including facial recognition data) with vaccination report credentials, the issue of surveillance by State and coercion countermeasures come up.

Like all things in the digital identity (DI) space this is not merely a topic where better technology will solve things. ‘Solutionism’ has often led to hasty deployment of poorly thought out designs causing no amount of disruption. There are real requirements to consider open standards based interoperability because it is almost comical to imagine that there is not going to be any adventures in silo making.

A key conceptual construct is to think about the number of overlapping or intersecting ecosystems which are positioned as stakeholders when digitally verifiable vaccination records become widely available. Almost every single business vertical or line of business is conceivably its own unique ecosystem. This means that these clusters of ecosystems will have their unique sets of rules - business, legal and technical which are applied to data exchange routines within the systems. And when a member of the ecosystem has to transition to another, rules of transitive trust have to kick in to enable recognition of assertions made by any holder of the credential.

The fun bit is that this is not complex. An implementation of this already exists in real life when various systems exchange paper based credentials across ecosystem boundaries. To be able to transform this into a seamless machine-readable contract based system would need a focused and concerted effort around testing. Rolling back hundreds of thousands of issued credentials through revocation and reissuing might be inexpensive (in money terms) but very expensive when viewed in terms of trust. Consider the situation if any nation had to reissue the entire set of active identity document issued to citizens. In terms of disruption to business and uncertainty it brings about it is unimaginable.

The only way to evaluate emerging solutions in the domain of vaccination reports is to examine whether they allow the recipient/holder

  • to have privacy and security

  • to be able to govern the credential and associated identifiers

  • to be able to optionally share specific parts of the data

  • to be able to use it without requiring additional special infrastructure

It is quite likely that at least one of the above conditions will not be met - design of VCs are still intimately tied to the presence of a ‘digital wallet’ app on a smartphone. However, it is also worth considering that there are various methods being designed to address this gap and make it possible to create a form of ‘low tech’ verifiable credentials.