The COVID-19 Credentials Initiative (CCI):
Bringing emerging privacy-preserving technology to a public health crisis
We submitted this position statement to the “Privacy & Pandemics Workshop: Responsible Uses of Technology and Health Data During Times of Crisis — An International Tech and Data Conference” by the Future of Privacy Forum. The aim is to, from two participants’ point of view, share an abbreviated case study of CCI and to highlight key challenges that arose in our efforts to responsibly use new privacy-preserving technologies to mitigate the spread of COVID-19. We wanted to share our submission with the CCI community and the public in the hope to invite some further discussions.
So what is this emerging technology that is just coming to market with very early adopters and innovators leading the way (including the US government, the government of British Columbia, etc.)? It is called Verifiable Credentials (VC or VCs), a data format standard developed and published by the W3C last year. It is a universal data format for one entity (person, organization or thing) to assert something about another entity. The issuer packages up the credential and cryptographically signs it to seal the data it contains. It passes this to the subject, or holder, so that the subject can share it with the receiver of the credential, the verifier. At that point, the receiver is able to use cryptography to check the seal and the validity of the issuer. VCs are privacy-preserving because the issuer (identity provider) and verifier (relying party) do not form a technical/federated link with each other. The information does not pass directly.
This technology gives people the ability to collect and manage digital credentials — similar to the cards we find in our physical wallets — in digital wallets. These digital credentials act like paper credentials because individuals do not need to have the verifier directly connect to the issuer. Why is this new? Until now, to have provable information exchanged online, the issuer and verifier would need to directly “federate” to exchange information about the data subject.
Immunology is a complex science, but simplified basics about how it works with some viruses are known. It was based on these simple understandings that possibilities for how VCs might be used by people and institutions to better manage risk began to be explored. One obvious use case was the ability to issue VCs that reflect some type of COVID-19 status, a proof that one:
- Tests positive for antibodies and therefore not infected/or vulnerable to being infected.
- Has recently tested negative and therefore the risk of being infected is low and that one could go to work, travel or visit a facility with vulnerable populations.
- Has received a vaccine for COVID-19 and therefore safe to travel or access a large in-person event.
Some of these use cases as articulated are not new (e.g. the Yellow Card); it’s just new to digitize them. It makes sense to consider how this simple paper-based technology can be updated to digital. However, the mild hysteria raised about doing things today done on paper via digital means that concerns were triggered without fully understanding or exploring how to use the technology. Some individuals were so concerned that they resigned from organizations whose leadership floated these ideas.
COVID-19 status is currently shared in two forms: 1) with the patient via a phone call or text message from the doctor or testing site or 2) in a patient medical record. Neither of these solutions provides a clear way for the subject to prove results to an entity that wants to know this information (e.g. an airline). VCs offer a new innovative format that provides people with information about their COVID-19 status that is under their control and verifiable by a relying party.
The VC approach is in strong contrast to the CommonPass effort, led by a Rockefeller-Foundation-backed nonprofit, that co-arose in the same time frame. The leadership at CommonPass is connected to the conventional medical records world and proposed the creation of a global system where patient medical records in some yet-to-be-determined way would be shared with a centralized decision engine that CommonPass would run globally. They held several global meetings with hundreds of people, including government leaders attending to build momentum for their proposed solution. There are also scores of siloed proprietary solutions popping up to solve these data-sharing challenges. For example, CLEAR is offering a biometric data sharing solution. In a public health crisis, these non-interoperable solutions are only good when only one of them is widely adopted, which doesn’t seem to be the case.
The community that formed around CCI was mostly made up of small, early-stage startups who were already implementing VCs for other use cases and decided to collaborate on exploring COVID-19 use cases. This makeup of the community means that it is not connected to global elites, governments, health departments or healthcare institutions. One exception to this is a startup that had political connections and worked with a California State Legislator to have the bill AB2004 proposed, which opened up a committee to study the use of VCs for COVID-19 medical test results. The bill just passed the Senate and is now waiting to be signed by the governor to become a law. However, the Electronic Frontier Foundation has repeatedly voiced its opposition  .
CCI has struggled to “raise voice” and be “heard” by the powers that make decisions. The experience of this group raises questions:
- How can emerging technology be “seen” by actors (governments, public health officials, airlines, workplaces) in the marketplace looking for solutions? Which technologists are listened to by policymakers? How are these actors making decisions about the claims these technologists are making?
- How can networks of potential issuers of COVID-19 status credentials be spun up in such a way that the credentials issued are seen as valid by verifiers (airlines, etc.)? This set of challenges around the technology are not technical as much as they are about process and accountability systems.
- How can existing norms of information sharing about people that are paper-based today be translated into a digital form in ways that make themselves and those concerned about human rights implications comfortable with the technical deployment?
- How can a new privacy-preserving technology based on open standards be used for public health crises or any other time-sensitive occasions when its deployment and adoption require a lot of coordination, collaboration and communication? How can these things be facilitated by funders seeking to make a difference?
We hope to hear your thoughts on the raised questions.