Spotlight on ‘Lawtech’: how machine learning is disrupting the legal sector
Ask a lawyer to describe their daily work, and you’ll probably hear the words ‘document-intensive’ in response. It’s an apt description. The legal profession is overburdened with paperwork. Legal rulings, past cases, contracts and myriad other papers contribute to a data proliferation that is hard to keep on top of.
Unfortunately, failing to keep on top of it is not an option. The thankless task of sifting, reviewing and summarizing often falls to the lot of new associates, who spend between 31 and 35 percent of their time conducting research, according to ‘New Attorney Research Methods Survey’, a study from the Research Intelligence Group.
It’s costly, time-consuming, and (probably) extremely dull. But this way of working is beginning to change. A new sector is emerging in response to the ever-growing burden of data and the resulting high price points for clients. ‘Lawtech’ is bridging the gap between this most traditional of professions and a hitherto untapped source of help: cognitive computing, machine learning, and artificial intelligence.
The convergence point: human expertise meets AI
Over the course of this blog, we’ll look at two broad use cases for cognitive computing, machine learning and artificial intelligence in the legal profession:
- Technology-aided review
- Smart document creation
Let’s be clear: the idea is not to replace human lawyers with search bots, but to set them free to do more interesting and high-level work by eliminating the miserable grind. Artificial intelligence is not good at writing briefs, appearing in court, or negotiating with clients. This work remains the proviso of highly skilled human beings, and rightly so. However, it is good at document review and data extraction, and has the potential to realize enormous benefits for the legal sector.
Let’s look at some use cases.
Technology-aided review (TAR)
Cognitive computing, artificial intelligence and machine learning are three pillars of something known as TAR – or technology-aided review. This is the process of extracting relevant data points from unstructured data sets, like legal documents or contracts.
It’s in the ability to work with unstructured data sets that AI and cognitive-powered tools come into their own. While it is relatively simple to mine a nice, orderly spreadsheet for information, it’s much more difficult to extract it from more convoluted or scattered sources. TAR tools are capable of ingesting vast data sets of any specification – including data from the Internet of Things. Even more significant, perhaps, is their capacity to learn. These tools go far beyond just programming – instead, they refine their skills and knowledge with each interaction or query, developing and learning as they go.
One such tool is ROSS Intelligence, built on the Watson cognitive computing platform. Its natural language understanding capabilities mean it can understand and interpret nuanced questions more accurately than a simple keyword search. To answer a question, ROSS sifts through multiple text documents until it comes up with the relevant information. The volume of data it can handle is staggering: up to a billion text documents per second. The sheer speed means huge benefits in terms of reduced labor and costs.
TAR tools like ROSS have multiple use cases. They could be invaluable for consultancy work – extracting data such as contract start and end points, or payment dates, from large numbers of documents. The tool then presents these findings in a handy dashboard form, where it provides the base point for human analysis, theorizing or negotiation.
It could also be used to analyze target companies in mergers and acquisitions deals. Machine learning can perform searches on particular companies and identify wording that differs from the norm in global sales contracts, to spotlight potential new deals.
Smart contracts and document creation
While research and review is the obvious use case, automation through machine learning has the potential to perform more complex tasks, too. One of these is document generation.
Contract generation software enables contracts to be automatically produced by asking a series of questions, such as: ‘when does the agreement begin?’ or ‘is the tenant undertaking work on the property?’ The questions generate a decision tree that determines the form the contract will take, ensuring all bases are covered. It’s a little like setting up formulae in an excel spreadsheet, to determine the rules by which the content therein will be governed.
Specialized document software of this kind can also enhance document organization by ensuring that all internal cross-references apply language consistently – even if multiple people have been involved in drafting them. This means consistent terminology across the document and minimized risk of misinterpretation. Document comparison tools can also check for undefined terms and identify missing conditions or clauses, to ensure a water-tight result.
Interestingly, the demand for solutions like these is creating a brand new ‘Lawtech’ sector, and new jobs with it. ‘Legal technicians’, or ‘legal engineers’, many of them ex lawyers themselves, have the job of designing automation solutions and helping law firms to implement them, without the need for a centralized IT system.
Software operating on a cloud-based platform is agile enough to grow and adapt to firms’ growing needs. Cognitive computing and artificial intelligence means the platform is ever-learning, self-improving and immune to the threat of becoming obsolete, because it is always updating its knowledge.
The future of ‘Lawtech’
The emergence of ‘Lawtech’ and its accompanying job opportunities marks an interesting convergence point between mankind and machine. Large organizations that have invested in automation software are already seeing significant returns, in the shape of reduced operating costs, swift, accurate data extraction and better opportunities for their staff to take on high level work.
Attorneys have more time to engage in the creative side of legal representation – keeping clients better informed throughout the legal process, and exploring strategies and outcomes fully with the benefit of trustworthy data. The value of cognitive computing in the legal sector is already apparent – and it’s not going anywhere.