Like many organisations, law firms often have to redact information from legal documents, such as within DSARs (Data Subject Access Requests). Whilst acting for clients involved in litigation, documents may need to be redacted to remove references to individuals other than the claimant (or others) in the action or to remove sensitive information (e.g. in medical or social work records). In both cases this is a time intensive task that is not very cost-effective for clients and can be less satisfying for solicitors than more complex legal case work.
Weightmans decided to find out if they could use artificial intelligence (AI) technologies to speed up document redaction. Funded by STFC's Bridging for Innovators programme, the company worked with data scientists at the Hartree Centre to develop a demonstration model to test what was possible using AI technologies like natural language processing (NLP) to automatically detect and identify words or phrases in documents that solicitors would have to look for, such as names, dates or locations. Due to the sensitivity of this kind of data, the scientists were unable to use real or “live” data so they used open data examples.
Redaction also poses challenges in itself, as it can inadvertently render information unusable unless reasonable replacement terms are added to the report to provide context – a page with everything removed isn’t much use!
The model performed well against current standards, demonstrating the potential of AI technologies to benefit law firms by streamlining their ability to pick out certain sensitive information and prepare it for redaction. The technology has the potential to enhance accuracy and reduce the time it takes to complete document redaction, freeing up lawyers to work on tasks where they can best use their knowledge and expertise to add value for clients.
This also has the potential to provide greater job satisfaction for individuals and teams, while faster document processing at reduced cost would provide financial benefits to clients. This project team will now explore routes to further develop and commercialise the model.
“This project is one way we are using technology to assist with tasks, without replacing the skill and expertise of our lawyers. Some tasks might seem mundane on the surface but they are incredibly significant for our clients – for example in a childcare case, it could be catastrophic to leave family contact details in. Using AI technologies can assist with efficiency of redaction, whilst also ensuring we maintain our high professional standards. By having AI technologies assist our legal experts we get the best of both worlds.”