Travers Smith LLP has today announced the launch of Etatonna, an innovative contract labelling tool, which has been developed in-house by the firm's Legal Technology team. Having successfully launched the tool internally, the firm has now open sourced the code, allowing others to use Etatonna within their businesses and benefit from being able to efficiently label legal documents to train AI models.
Etatonna makes it possible to easily label concepts in contracts for the purpose of automated document review. Labelling data in Etatonna is stored in a readable structured format to easily train AI models. Once the data is captured, it can be used to generate, and regenerate, machine learning models for any purpose, simply by sending the structured data to any third-party system API to auto-generate models.
Using Etatonna, businesses have full autonomy over what data exists within machine learning models allowing them to retain their investment in data and training and even share datasets to build joint venture-type AI models.
During the development of Etatonna it became clear that certain legal documents lend themselves to being labelled by non-experts. For example, clear headings and short paragraphs, in almost all cases, will provide enough context for non-experts to label correctly. This means that users do not need to rely on legally trained experts in order to use the software efficiently.
The introduction of Etatonna provides a solution to one of the biggest problems in the legal industry, ownership of the IP of trained data models. By open sourcing the base platform, Travers Smith has made it easier for businesses to create their own datasets and share them in their original form, while still retaining the IP.
Head of Legal Technology, Shawn Curran, said: "There is no doubt that using artificial intelligence to unlock insights from historic contracts is of immense value to organisations all over the world. The main issue at the moment is the quality of the AI. Therefore, model training needs to be optimised and companies shouldn’t be overlapping by labelling the same or similar clauses. Law firms need to combine expertise in the most efficient and cost-effective way for our clients. Our hope is that by open sourcing Etatonna we can inspire this collaboration."
Sam Lansley, Software Engineer, said: “The lack of transparency with machine learning models is a big concern for most industries. Models are simply a combination of "learned" algorithms and at the moment, this makes it extremely difficult to track them back to their source data. Etatonna completely solves this problem by linking the model back to the training data, and then linking that back to the original documents.”
Organisations that would like to access the open source code can register by emailing Shawn Curran at shawn.curran@traverssmith.com. Etatonna is licenced under the GNU GPL v3 licence and is shared through the existing Travers Smith Open Source environment on Azure DevOps.
In March 2020, Travers Smith launched the first product from the Etatonna platform, a free Force Majeure app. The app reviewed contracts for Force Majeure provisions and once identified, interrogated the clause for key terms like pandemic wording or a termination right.
Since joining the firm in 2019, Shawn Curran has built up a legal technology team at the firm that brings together engineers and lawyers to develop products to solve some of the biggest challenges facing the legal industry. In October 2019, the team launched its open source MatMail software, which significantly reduces time spent on filing emails. Since then the team has also developed two apps as part of the FT's Global Legal Hackathon, which aimed to solve problems caused by the Covid-19 pandemic.
