Artificial intelligence doesn’t mean the end of lawyers – at least not yet

Elizabeth Huang

‘Will robots replace lawyers?’ seems to be the question on the lips of every legal commentator looking to weigh in on the debate surrounding the role of artificial intelligence (AI) in the legal industry. Major media outlets such as The Guardian have covered the issue, claiming dramatically (among other things) that judges have a 40% chance of being replaced by ‘robots’. Cue hysteria, as panicked law students reconsider (for the nth time) their choice of degree.

Not so fast.

After all, it takes a lot to replace the human touch in an industry such as law, which requires creativity, empathy and a whole host of other complex skills. The current framing of the AI debate, with its dramatic language of replacement and redundancy, diverts attention from the much more pressing question of how lawyers can adapt to, and incorporate, AI technologies into their current work processes. For now at least, Lawtech is friend, not foe. We would do better to consider how its functionality can be maximised, rather than regarding it with the suspicious gaze of a competitor.

What’s out there at the moment?

One way of categorising the tools currently available is to think about where they might fit within the lifecycle of a legal issue or dispute. From initial advice to the final payment of fees, Lawtech can play a role at every stage of the process. Lawtech is a hugely innovative area: as much as £16m has already been invested in start-ups, as funders recognise the industry’s potential for rapid growth.


Lawbots are typically customer-facing legal AI applications. They can be used to automate responses to client enquiries, identify legal issues or provide straightforward legal advice. A well-known example is DoNotPay, which assists individuals by automating responses to simple legal issues such as parking tickets and tenancy contract violations. Last year, a group of Cambridge law students successfully developed their own chatbot, and have now gone on to found a lawtech start-up called CaseCrunch. Lawbots may guide clients through the entire legal process (where the process itself is extremely simple), or assist them by providing initial legal advice (where the client may be unclear as to whether there is a legal issue worth pursuing at all). At the moment, there is a gap in the market for a lawbot able to join these two tasks together and provide a whole-service package to clients.

Fundamentally, the issue is that the kinds of problems for which clients require initial legal advice tend to be inherently more complex, and would require legal argumentation and processing beyond the current capacities of lawbots. If a lawbot could be created that was able to provide both advice and execution, lawyers might need to seriously start reconsidering their career options…

Case outcome prediction

Case outcome prediction software is also used at the initial stages of the legal process and may help clients decide whether a particular issue is worth pursuing. Across the Atlantic, Docket Alarm, a company that specialises in the field of patent law, enables lawyers and clients to track relevant court cases and use analytics to predict case outcomes (based on the decisions made in previous cases). Factors that can be taken into consideration include the judge attached to the case, the law firm instructed and the type of technology under dispute. More controversially, courts in the US use algorithms such as COMPAS to assess the future risks posed by individual offenders. The exact algorithm is confidential, and its application raises difficult questions about the ethical issues linked to the use of technology in this way. Case outcome prediction software is an excellent example of how ‘big data’ can be utilised: increasingly powerful analytics enable patterns to be identified in legal decision-making.

Document analysis and review

This covers a broad range of applications including legal research, automated document generation and due diligence review. RAVN, used by Linklaters and Berwin Leighton Paisner, is able to read, interpret and extract information from documents at greater speed and accuracy than human lawyers. Another example is Kira, which specialises in contract clause analysis and review, with a focus on improving compliance – among its users are Freshfields Bruckhaus Deringer, Clifford Chance and Herbert Smith Freehills.  Case management software improves efficiency and produces higher-quality, higher-accuracy documents.

The short to medium-term impacts

Three key impacts can be identified. Firstly, the use of AI technologies has clear implications for legal roles such as those of paralegals and trainee solicitors. Administrative work is the category of work most likely to be automated in future. While paralegals may find themselves in trouble, trainees may benefit from the opportunity to take on more sophisticated work, as they will be freed from the drudgery of filling in forms and proofreading documents.

Secondly, the relationship between lawyers and clients may also evolve. If clients find themselves able to consult cheaper and faster lawbot-type programmes before deciding whether they need to consult a human lawyer, firms may discover that they need to become better at actively engaging with clients and providing advice through more efficient and innovative means. One likely possibility is that clients will take on a more assertive and proactive role, forcing firms to adapt accordingly and respond to increasingly well-informed and opinionated clients.

Finally, there are also implications for the way in which firms operate and bill clients. The automation of workflows and processes will inevitably lead to increasing pressure (asserted by clients) for greater efficiency and subsequently lower fees. In the face of this, more and more firms are moving towards non-traditional business models e.g. fixed fees, freelance working styles and alternative business structures (ABS), which lie somewhere between in-house lawyers and professional services.

No need to panic yet…

"Let’s kill all the lawyers", declares Dick the Butcher (or rather, Shakespeare), in Henry VI. Are we nearing a future in which, Terminator-like, Lawtech declares the same: war against lawyers – man vs. machine? The short answer is no.

Most simply, the technology is not yet sufficiently advanced to replace the work of human lawyers. What’s more, on the latest estimates, the work most likely to be automated comprises only a small percentage (around 5%) of lawyers’ overall billing.

Even in what we might think of as the ‘worst-case scenario’, in which AI becomes as good (if not, inevitably, better) as human lawyers, this does not necessarily mean the end of the legal profession. AI tools would still require monitoring and scrutiny – work that would need to be done by people.

The world has already been through several technological revolutions - lawyers have survived and thrived in a changing legal landscape, each time relying on their distinctive human skills. We are not yet at a point where Lawtech can even come close to approximating these so for now at least, lawyers can breathe a sigh of relief. Counter-intuitive or ironic as it may seem, it is the legal industry’s humanity that may well be its saving grace in this brave new world of artificial intelligence.