Peter Hunn, the founder of Clause, has launched Legal Schema – an ambitious project to create a ‘common framework for the expression of structured data in contracts’. The project is being supported through the Lawtech Sandbox.
It’s also another example of the growing trend toward developing operational standards across the legal and legal tech world.
This naturally will be of benefit to smart contract makers, but also to anyone working with contract data, e.g. when using a contract automation tool and then seeking to extract key information from that document.
As Hunn explained to Artificial Lawyer: ‘Legal Schema will enable legal documents to be easily structured and consumed as data. You need to structure data in a common way or you lose the value of the data.’
I.e. if we are building contracts digitally, and we then want to extract, store and share key aspects of the contract’s data, e.g. names, dates, addresses, jurisdictions, key provision terms and more, then it will help the market as a whole if everyone is using the same underlying definitions in this digital environment.
Artificial Lawyer asked if this is meant to be a kind of common HTML for legal data? Hunn replied: ‘It’s beyond that, this is a structured schema, such as you see with Schema.org, which provides a standard way to describe data on the internet (N.B. that is not Hunn’s project, his is Legal Schema).’
‘[You need this when using the internet] so Google can know what ‘movie’ is, for example, wherever it is mentioned.’
Hunn said that they would continue to work with the Lawtech Sandbox, which has been created by the Lawtech UK group, headed by Tech Nation’s Jenifer Swallow. They will also be reaching out to UK legal bodies and academic institutions to help with the project.
Hunn, as with an increasing number of legal tech experts, is a big proponent of the structure first school of thought when it comes to contracts.
‘Today, a contract is just words on a page, there is no structure to it. Later, NLP tools are used to make structure from this, but if you start with structure to begin with then you don’t need NLP,’ he explained.
(Although….of course, there are today literally billions of contracts in existence that have not been made in a highly structured way and will still need NLP. And most contracts are still made in a highly unstructured way too, at least for now.)
The structured data approach also then creates a new problem: how is everyone classifying and defining the data they isolate and extract using digital tools? Can that structured data then be shared with others, and with other digital tools, that will be using the same underlying definitions?
This is especially true of smart contracts. If a firm makes a smart contract that has coded elements related to key terms in the contract, how can another party engage with that contract digitally if they are not necessarily using the same definitions or ‘data styling’ to describe what has been shared?
That said, Hunn was at pains to point out that this is not all about smart contracts that self-execute, but is a more ‘back to basics’ approach around making digital contracts in general more effective.
This will matter to a far broader audience, he explained, because when ‘you move away from just having text [to describe what’s in a contract] you need a structured representation so you can then query that data more easily’.
In conclusion, Hunn added that the idea here is to create a public good, that is a benefit to everyone.
Other companies in the Lawtech Sandbox include: Amplified Global, Legal Utopia, Clausematch and Deriskly. They have all been part of the initial pilot stage of the sandbox which took place over a three month period.
During this time they all saw ‘accelerated growth, realised through a number of tailored tools and services, including: access to regulators for fast response support in navigating regulation and governance; connections to LawtechUK’s network to help drive collaboration, commercial opportunities and development of legal tech products; and access to legal data by matchmaking with data sources’ LawtechUK commented.
[ Main pic: shows a screenshot of part of an example from Schema.org for ‘legislation’. As mentioned, this is not Hunn’s project, his will be Legal Schema with a focus on contract language, and the example is used here for illustrative purposes only.]