“Until you know what it does, it’s artificial intelligence (AI); then it’s just software” is a good way to demystify AI and works just as well for law firm use of AI as for any other sector. Almost halfway through 2016, AI software in the legal services market looks like it is really starting to take off, with firms announcing in increasing numbers that they’re “getting an AI”.

2015: some straws in the wind

Global law firm Dentons’ NextLaw Labs reported in August that it had invested in ROSS Intelligence, the IBM Watson legal advisor app developed by the University of Toronto. In September, international firm BLP announced that it was using RAVN Systems AI platform ACE (Applied Cognitive Engine) to develop LONald, a property contract robot for issuing light obstruction notices (LONs).

In October, Thomson Reuters (TR) confirmed that it had entered into a long term agreement with IBM to use Watson technology in product development across all its businesses. TR subsequently confirmed that TR Legal would be the company’s first division to release a product with Watson “under the hood”.

May 2016: important developments in the US and UK

On 5 May, AmLaw 100 firm Baker & Hostetler became the first US law firm to announce that it had licensed ROSS, saying that it would be used in the firm’s bankruptcy practice, which is acting for the bankruptcy trustee in the liquidation of Madoff Investment Securities.

UK Magic Circle firm Linklaters confirmed later the same month that it had entered into a Master Services Agreement with RAVN Systems for its ACE AI platform. No details have as yet been given about the particular applications, services or activities that ACE will be used for.

On 20 May, at the CodeX Future Law Conference at Stanford University in California, Andrew Arruda, ROSS Intelligence CEO, announced further ROSS licensing deals with international firm Latham & Watkins and Milwaukee based von Briesen & Roper.

Watson and ROSS have four main benefits:

  • A natural language user interface that enables everyone to type or speak their question in plain English. (For example, “What’s the earliest date I can terminate this contract”?).
  • Awareness of the context of the question, so it can better analyse large datasets (like legislation, cases and in-house and online know-how).
  • Generation of plain English evidence-based responses ranked from most to least likely.
  • Most importantly, AI is cognitive: it learns from user feedback so that the next time it is asked the same question its answer will be better.

It’s these benefits that Dentons, Baker & Hostetler, Latham & Watkins and von Briesen & Roper are buying into for their own clients, and that TR are developing for their products and services for their law firm and in-house customers.

Predictive coding software

BLP has recently became the first UK law firm to win a High Court ruling ordering the use of predictive coding software as part of the litigation discovery document review process. Predictive coding is another form of AI and, like Watson’s ROSS and RAVN’s ACE, uses machine learning to identify electronically stored information through context awareness and cognition.

It works by lawyers on the litigation team iteratively “training” the software in what to look for in a “seed set” of relevant documents, providing feedback to tune for accuracy. The software then applies the training it has learned from the seed set to the whole dataset of documents under review.

The take-up of ROSS, ACE and predictive coding shows just how quickly legal AI is gaining traction in the UK and US markets, and demonstrates how fast “computerisable” chunks of even high-value work, which were previously carried out by lawyers, are starting to be commoditised and replaced by “as a service” provision.

Mass market AI developments

Hand-in-hand with legal sector-specific applications go the mass market AI developments of the tech majors: Siri from Apple, Cortana and Bing from Microsoft, Echo from Amazon and Google Now. For example, Microsoft’s Cortana personal digital assistant on Windows 10 will learn to manage your diary, recognise your voice, set reminders and answer questions. Bing’s translator is also getting pretty good at instantly translating web pages and other larger amounts of text.

All these developments in law firm AI, both sector-specific and mass market, are contributing to a quickening rate of change in a legal services market that is increasingly driven by IT. As with the cognition attribute of AI that learns from user feedback to speed up response time, we can expect the rate of change to get faster.

First published on Practical Law In-House Blog.