♫ Science fiction double feature – Doctor X will build a creature
See androids fighting Brad and Janet – Anne Francis stars in Forbidden Planet
Wo oh oh oh oh oh – At the late night, double feature, picture show... ♫
- Music and Lyrics by Richard O'Brien from The Rocky Horror Picture Show
There was a time – not long ago – when the application of artificial intelligence (“AI”) to the practice of law was considered, well, science fiction. However, AI is no longer seated in the back row – it seems worlds have collided. The future of legal contract review is now and has been created in the lab, not by lawyers but by engineers and mathematicians (at least in part).
Legal contract review (“LCR”) incorporates and is dependent on several aspects of AI. Machine learning, one such aspect, evolved from pattern recognition by the construction of algorithms that can learn from and make predictions on data (Wikipedia), and is closely related to computational statistics. When the data set comprises a collection of legal contracts, LCR sorts documents, finds key provisions and spots critical provisions. How effective is it? “A study of Diligen (a LCR platform – editor) customers over a 12-month period revealed that the platform increases the accuracy of contract analysis by 80% compared to a lawyer manually reviewing contracts alone.” (The Impact of Artificial Intelligence on Law Firms).
Another aspect of LCR is natural language processing (“NLP”). NLP is also built on machine learning algorithms and is based on statistical models that increasingly rely on deep learning techniques. Lest you think that statistical models are a far cry from real human word analysis, according to Wikipedia, deep neural networks have produced results in NLP and other fields “comparable to and in some cases, superior to human experts.” NLP has been around since the early 2000s. It uses deep learning to translate an atomic word into a mathematical positional representation of the word relative to other words represented as a point in a vector space. All clear?
Want to see it at work? Google Translate uses NLP to “translate whole sentences at a time, rather than pieces.” It learns from millions of examples and “encodes the semantics of the sentence rather than simply memorizing phrase-to-phrase translations.” Bit of an improvement over the old English to Italian translation books, eh? Another application is the speech recognition that we are seeing pop up in so many contexts on the web and incorporated into human-machine interface devices today (Apple’s Siri uses Nuance’s speech recognition engine – the same one that was developed for, and which lawyers now use in Dragon Dictate).
You can glimpse signs that AI is changing things. iManage (the document management platform used by many large and medium law firms) in May 2017 acquired RAVN Systems, an artificial intelligence platform that allows users to organize, discover and summarize documents.
Blue Hill Research published a Benchmark Report in January 2017 to evaluate ROSS Intelligence and AI in legal research (a closely aligned area of AI to contract review). ROSS is an advanced NLP legal research tool built on IBM’s Watson AI cognitive computing platform. It can produce legal memoranda and monitor the law on specified legal issues in the areas of Bankruptcy Law, Intellectual Property and Labour & Employment (US law only, at this point).
Blue Hill found that the ROSS AI tool provided a significant, additive contribution to the effectiveness of legal researchers, resulting in a reduction between 22.3% and 30.3% in research time compared to groups using solely Boolean or natural language searches. Users also reported higher levels of confidence and satisfaction in results using the ROSS tool.
It seems that Doctor X has built a creature that is transforming the legal landscape, one artificial bit of intelligence at a time. It is allowing lawyers “to analyze legal issues and make connections that would otherwise be invisible.” Talk about science fiction – here is the invisible man!