Nine months ago, the University of Surrey in collaboration with the Open Data Institute and the UK National Archives began work on blockchain-based research project, ARCHANGEL. Funded by the UK Research Council, the project is an attempt to use blockchain technology to make historical records resistant to alteration.
Originally planned to last 18 months, the project is now expected to span two years.
By creating a hash of the digital contents of an electronic record, the data can be placed on a blockchain, where changes can be tracked. While this doesn't solve the problem of a falsified source document, it can verify that a record has not been altered while under the care of the archive itself, starting from the time the document's hash is uploaded to the blockchain.
"We're particularly interested in situations where the archive has information that is currently closed [to the general public]. We put the hash on the blockchain today. And then … when the information can be made available, people will be able to verify – when they see the open information – that during the period of time it was held by the archive, it's not been changed," said Jim Sheridan, digital director of The National Archives in a recent interview with ETHNews.
The involved parties are considering whether to establish ARCHANGEL as a permissioned blockchain, which would mean that only participating national archives, universities, and libraries would be able to upload information, though the information on the blockchain would be publicly viewable.
Sheridan explained that such a system would avoid a major pitfall of public blockchains: "We think one of the benefits of a permissioned blockchain would be that it wouldn't suffer from this recentralization happening on the general blockchain because of the economy of scale and compute power."
According to Sheridan, the consensus algorithm would be like Bitcoin's, requiring more than half of participating institutions to validate any changes made to the blockchain: The more participating institutions, the more secure the blockchain.
For this reason, ARCHANGEL is envisioned as a multi-institutional, and possibly even international blockchain. The archival institution hopes to attract additional partners by fall.
"I can't name names, but we do have several other memory institutions keen to work with us," said Sheridan.
During this early test phase, The National Archives is focusing on using ARCHANGEL to archive born-digital records, particularly research data and videos of UK Supreme Court proceedings, and the project is expected to expand.
"We think that those are interesting and important use cases. If you can do it for those things, which are reasonably difficult, then things like email or standard office documents … should be comparatively easy," Sheridan explained.
While a National Archives blog post on the project claimed existing digital technologies can allow someone to "rewrite history," Sheridan maintains that the project is not an attempt to solve some present problem of rampant historical falsification – a practice he admits is not widespread. Rather, it's an attempt to maintain the level of trust that citizens currently have in archives and memory institutions.
But even if the project is found to be feasible and all the technical issues are solved, its success hinges on whether this technological solution is appropriate for what is essentially a psychological problem: trust.
"It's about maintaining the high level of trust that we have, into the future. And into a future where the level of trust that people will have, particularly about digital evidence of all forms, is going to be continually eroded."
However, the additional work the archive is proposing can only lead to ongoing trust if the general public understands the technology and can be convinced (that is, can trust) it is as tamper-proof as Sheridan claims. Currently, blockchain technology remains relatively obscure.
"We know that most people don't understand what a blockchain is, so saying 'hey our records are on a blockchain, so you can trust them,' doesn't mean very much to many people who are using our collection," he says.
"We've got to do some experimentation about how we best convey this to users."