AI, A New Tool For A New Type Of Lawyer – ThoughtRiver

This is a Sponsored Thought Leadership article by Ben Green, and Tim Pullan, Founder and CEO of legal AI company, ThoughtRiver.


Since the turn of the millennium, technology has changed the way we work beyond recognition. Jobs, which a generation ago required manual labour are now performed by machines that are several times faster, do not take sick leave and need no rest. The revolutionary effects of technology do not extend merely to blue-collar jobs; now, even banks, hospitals and supermarkets are utilizing technology to best serve you as a customer, while maintaining their margins.

There are some industries which have staved off most of the impacts of technology so far – legal being the most notable of these. This is not to say the legal profession is staffed with Luddites who oppose the tides of futurization. After all, lawyers were quick to see the value of online research databases, which facilitate quicker access to case law and legislation. Rather, the legal profession is noteworthy for being apparently ‘unfit’ for the changes brought by the latest wave of technological advancements: namely artificial intelligence.

The Technology

Tim Pullan, CEO, ThoughtRiver

Artificial intelligence (A.I.) is complex, yet becoming increasingly common in everyday life. Google, Apple and Amazon have all invested heavily in A.I. technology, which is now at the fingertips of almost everyone in the world with access to digital devices and connected to the internet. Most types of A.I. today is enabled by machine learning, which is a technique that allows A.I. to ‘learn without being explicitly programmed’ (Arthur Samuel 1959).

Machine learning is used by Google’s translation tool to recognise meaning and grammar patterns within sentences, Amazon and Facebook to target advertising specific to your interests, as well as Apple’s intelligent personal assistant Siri to answer specific questions and make recommendations.

However, not all service applications are suitable for this technology: healthcare workers, teachers and counsellors require a significant human touch in the tasks they perform and A.I. cannot simulate empathy and compassion, yet. The essentially human nature of these professions makes them immune to the advance in A.I. technology.

The Profession

Lawyers often run the above argument, claiming that the personal nature of their clients’ legal problems requires empathy and a human touch, which machines simply cannot provide. This is true to a certain extent: the core aspects of a lawyer’s job are inescapably human: devising an effective commercial negotiation strategy, devising custom contract provisions, advising on highly nuanced and grey areas of regulation, or dealing in areas like family and employment law which require empathy.

However, there is a dichotomy to be made here between the more human aspects of a lawyer’s job and other more automatable tasks. Behind every piece of advice that a lawyer gives is a profusion of research, document review, drafting and due diligence – all repetitive and computable tasks. I.e. tasks that can be automated by A.I.

There may be some variation depending on a firm’s areas of practice and client base, but almost without exception, every lawyer’s work will be composed of both human and automatable tasks to some extent.

The Tools

In the past few years, there has been an infusion of legal technology that automates tasks thanks to the capabilities of A.I.

For example, ThoughtRiver has created Fathom, which is able to review contracts and analyse their potential risks. Fathom has been trained on thousands of contracts to recognize specific clauses (warranties, indemnities and exclusions of liability etc.) and employs machine learning algorithms to answer several legal questions, the type that lawyers themselves ask when reviewing contracts.

This type of technology can reach an unprecedented level of analytical depth in an increasing number of contract types. It will not just recognise the existence of a limitation of liability clause, but the scope of the limitation, and not just recognising a termination right, but also the conditions and notice period attached to that right.

The Result

Document review does not take up the entirety of a typical lawyer’s time, so technologies such as Fathom will not put solicitors on the brink of redundancy. Yet, the potential efficiency savings could be extraordinary. This is especially true within in-house departments where the passage of new contracts is near continuous: Fathom’s triaging capabilities allow general counsel to prioritise their time to focus on the riskiest contracts, safe in the knowledge that no contract goes without a risk analysis.

Lawyers are increasingly being required to think commercially and to develop more holistic solutions to legal problems. For now, A.I. can neither replicate the imagination, creativity or commercial nous of a good lawyer, nor negotiate or litigate on behalf of a client. Rather, legal technology automates the repetitive and delegable tasks.

This is reflected in Dana Remus and Frank Levy’s research paper, ‘Can Robots Be Lawyers? Computers, Lawyers, and the Practice of Law’. Remus and Levy indicated that the effects of A.I. would likely be concentrated on the more laborious tasks such as document review, drafting and case administration. These tasks combined make up 12.8 percent on average of billable hours compared with communicating, advising and acting for clients which makes up 35 percent (for more specialist contract lawyers, the document review element will typically comprise a higher percentage of their time).

Artificial intelligence is not a threat to the legal industry, but rather a tool to be exploited by lawyers for their own and their clients’ gain. Performing time consuming tasks in a fast, reliable and cost-efficient manner affords lawyers the time and knowledge they need to bring more creative and valuable solutions to the table for their clients.