Sponsored Content: Learning Machine

Remaking Credentials

min read

Learning Machine logo

As we move from a world of discrete paper repositories to a world of interconnected digital systems, we need academic records that are natively digital and self-verifying to reap the full benefits of electronic exchange.

With the rise of decentralized systems, blockchains have become famous for enabling a new level of security and peer-to-peer exchange for digital assets. Not to miss a marketing opportunity, some software vendors have added blockchain timestamping to their PDF credential service. This is a process by which a document is registered on a blockchain at a specific point in time to prove that a certain version existed at that time. The question here is, what added value does timestamping actually provide in this scenario? In reality, very little.

sample PDF digital credential shown as green rectangle
Figure 1. Academic records can be natively digital for greater functionality and convenience

Blockchains were made to enable a world where digital assets are cryptographically owned by recipients and can be freely shared and verified without relying on any vendor or third party. So unless a software provider has gone to lengths that make both of those goals real, no fundamental benefit is being realized from using a blockchain over traditional, digitally signed PDFs. When a vendor is still providing all of the assurances of document integrity, blockchain-based security is simply redundant. Further, proprietary approaches that aren't open source or based on open data standards are doomed to a short lifespan, as markets tend to coalesce around technologies that create the widest stakeholder ecosystems.

Even when a PDF has been digitally signed and timestamped on a blockchain, it doesn't suddenly become useful as a software object. You can view it and verify with reasonable certainty that a digital signature belonged to a particular party at the time of signature, but even this security has holes that have been exploited by fraudsters. PDFs were a kind of digital capstone or equivalent for the age of paper, which is evident because most institutions that issue PDF records still also issue the majority of their official records on paper. But PDFs are not the way to enter a truly digital age. We need to do better, and we now have the technology to do so. Official records can be software, made to interrelate with other systems in reliable and dynamic ways. This is how we achieve the automation, speed, analysis, and discovery that everyone desires.

Natively Digital Credentials

JSON has become the default file format for transmitting data on the internet and within web applications. While originally named for moving JavaScript objects, it is now used as the standard format for any popular programming language. The most common use cases are for web APIs that send data between third-party systems or to communicate within a system between a server and a user's browser.

As the de facto standard for transmitting web data, JSON should be the starting point for any type of official credential that seeks to take full advantage of the web and electronic exchange. This is why JSON was selected as the starting point for developing Blockcerts, the open standard for blockchain-based credentials, which Learning Machine launched with MIT in 2016. The primary question we had was how to fully equip a JSON file with the properties needed to operate as a modern credential. In addition to being instantly verifiable using a blockchain as a global notary, a few design principles were always priorities:

  • Open source
  • Reliance on open standards
  • Recipient ownership
  • Minimize resource requirements (computation, cost, etc.)
  • Must be viable without any proprietary product
  • Blockchain-agnostic

These minimal requirements resulted in a solution now regarded as the most secure, interoperable, and standards-based way to issue and verify natively digital records. Committed to eventual alignment with the W3C's Verifiable Credentials Specification, Blockcerts JSON files are digitally signed by an issuer and anchored to a blockchain for later verification. Even the visual presentation layer has been cryptographically sealed, so parties looking at the credential know all of the machine-readable data is fully integrated with what they are seeing on screen.

Further, each credential has an embedded cryptographic key unique to a recipient, allowing the recipient to prove ownership of the credential. This prevents use and reuse of the credential by unauthorized parties, which is important in an age of easily shareable digital records. The potential for computer systems to organize, filter, combine, and understand natively digital credentials is limitless—for systems used by both issuers and verifiers, or parties who rely on the credential to make a decision. Imagine having an HR system that automatically verified, organized, and used machine learning to help derive insight about a pool of applicants.

sample blockchain-secured digital transcript snapshot and on mobile device
Figure 2. Sample blockchain-secured digital transcript

Learning Machine's first product enabled schools to issue digital diplomas as Blockcerts, and today we work with research universities, community colleges, professional and online education providers, and K–12 campuses to issue blockchain-anchored diplomas. We have also recently launched the first product in the world for issuing blockchain-secured transcripts that are data rich, tamper proof, human readable, and machine readable. Learners only need to receive these documents once, and after that they can use them any time they are needed, such as when applying for a job. The confluence of the web, mobile technology, and blockchain has created the necessary infrastructure to provide transformative security and convenience for everyone involved, realizing the full promise of digital credential ownership, usability, and longevity.

In Summary

While PDF documents are digital, they carry many of the same limitations as paper. They are inert and heavy files whose value is confined mostly to their own display, which can't be fully parsed by computer systems requiring flawless document integrity and machine readability. It's not hard to see how the differences between PDFs and JSON files might get lost when they look very similar on a screen. However, the difference on the back-end is profound. Understanding that chasm starts with appreciating the full range of function inherent in software objects, including their ability to self-verify and their readiness to be ingested and parsed by other systems and processes.

Blockcerts, developed by MIT and Learning Machine, is the first realization of the full promise of digital records: now individuals can own their credentials for life, and anyone they share them with can verify them instantly and for free without relying on any vendor infrastructure. Educational institutions are already starting to transition to Blockcerts, many of them using the software Learning Machine has built to make it easy to issue and receive Blockcerts diplomas and transcripts. The future of academic records is here today.

Chris Jagers is CEO of Learning Machine.

© 2019 Chris Jagers