AI: hype or help?
Artificial intelligence has already shown it can help streamline processes but it's yet to change the way legal services are delivered â€“ that's the next challenge, says Sophie Brookes
Technology, and its speed of change, was the key challenge identified by law firms in PwC’s Annual Law Firms Survey 2018.
Artificial intelligence (AI), the simulation of human intelligence processes by machines, is nothing new.
In 1997 former world champion Garry Kasparov was beaten in a chess game by Deep Blue; in 2011 IBM’s Watson beat the two best human players of the game show Jeopardy; and in 2017 AlphaGo, a computer program trained to play the board game Go, beat the world’s best human player.
But it’s not just the gaming world that is feeling the impact of AI. Could it be that the machines are really taking over?
AI has existed in law firms for years but without the catchy acronym. Many time recording tools use AI to capture and record a lawyer’s activities, even interrogating the lawyer’s desktop and telephone footprint to identify the files she’s been working on.
Accounts systems automatically generate invoices at the end of each month based on the number of hours logged by fee earners.
Document management systems file and store documents according to where emails sent to the same person, or documents with the same parties, were previously stored.
The last five years has seen a boom in new AI tools or ‘legal tech’ aimed at automating different aspects of a lawyer’s work to make this faster and more accurate, and to streamline service delivery.
It seems the tech companies have looked at the way we work, considered us archaic and identified this as an area ripe for improvement.
Contract reviews, data extraction, document automation, proofreading tools, case strategy programmes and more are all flooding the market.
The challenge for law firms is seeing through the hype surrounding this wave of new technology to identify the solution which will help them most in practice.
In the scramble to be seen as innovative, adopting the wrong technology can be a costly mistake.
Upfront investment challenge
Many of the products require considerable upfront investment. But the finance model of a traditional law firm is not well-suited to incurring substantial costs where the benefits will be reaped over a number of future years.
Will the partners who incurred the cost still be there to benefit from the anticipated profits? DWF overcame this challenge in their recent IPO by raising £10m to fund improvements in IT.
Smaller firms may not have the same resources but they may benefit from a three-year project announced by the government that will see £3m spent on investigating how law (and insurance) businesses can make the best use of AI.
Another challenge for lawyers is simply understanding the technology in the first place.
Litigators might be good at defending claims, conveyancers can protect clients acquiring a property but do any of them have the skill set to analyse and understand complex new software?
While it’s not necessary to know exactly how the technology works, or how to write the underlying code, someone needs to invest the time to properly understand the benefits and capabilities of new innovations, translating them into a language and methodology which lawyers can easily grasp and embed into their working practices.
Increasingly, this is leading to the appointment of ‘innovation directors’ or ‘technology lawyers’ – someone whose specific role is to drive technological change within the firm.
Recognising that this is a different skill from those in which lawyers are traditionally trained, the University of Law now offers a degree in Law with Legal Innovation and Technology, and the University of Edinburgh offers a post-graduate course in Innovation, Technology and the Law.
When investigating new products it is possible to be simultaneously impressed by their capabilities but disappointed by their limitations.
Product X can be tailored to produce an efficient solution to problem Y; but it can’t provide a generic solution capable of applying across broadly similar, but specifically different, situations. Document reviews have historically suffered from this but that seems to be changing.
The Serious Fraud Office recently completed a four-year corruption investigation into Rolls-Royce.
It used Ravn, an AI document extraction tool, to sift through over 30 million documents, with the SFO’s chief technology officer Ben Denison saying in an interview with CIO that it would be impossible to investigate cases on this scale without using technology.
With such a range of possible solutions on the market, coupled with a limited budget, firms have to be selective in the tools they adopt.
It’s important not to get lost in the hype and rush into buying something simply because the demo looked good or because it worked for another firm.
Instead, they should carefully analyse existing practices to identify problems or pinch points; then look for something that will help resolve that specific issue, by doing it faster, cheaper or more accurately.
The implementation of new solutions will also have to be carefully managed to ensure the maximum benefit is achieved.
Recognising, and addressing, the need for a change management programme will be key to ensure new technology is properly embedded and utilised within the firm and doesn’t just become a costly vanity project.
Lawyers at risk?
While the business case for some products is still to be proved, document automation is one area which has gained real traction.
Whether it’s using inbuilt tools, like Practical Law’s FastDraft which automates documents available on their service, or using a managed service, like that provided by Clarilis to automate a firm’s own suite of precedents, the business case for automation stacks up. Clarilis claim to deliver an average 90 per cent reduction in drafting time.
And not only do lawyers benefit from this direct time saving, it also frees them up from handling low-value tasks to focus on more strategic work of real importance.
So are legal jobs at risk? It certainly seems likely that some tasks traditionally carried out by junior fee earners will be replaced by machines capable of completing them at a much higher rate.
In 2017 investment bank JP Morgan announced that it was using a program for contract intelligence to interpret commercial loan agreements, a task that historically had required up to 36,000 hours of lawyers’ time each year.
It’s hard to see how changes like this cannot have an impact on traditional jobs. But, rather than simply replacing humans, the rise of the machines has led to the creation of new roles within the industry – the legal technologists referred to above.
So it’s not so much that jobs will disappear, but that workforces will change. And simply using machines to more quickly and efficiently perform tasks usually done by lawyers fails to embrace the full potential of AI.
The real benefits will come by harnessing AI to fundamentally change the way legal services are delivered and experienced, with humans and machines working together seamlessly.
This is already happening on due diligence reviews with tools such as Kira or Luminence being used to flag those high-risk areas on which lawyers then need to focus their time.
Eversheds Sutherland says that it uses ThoughtRiver as “a way of triaging contracts and prioritizing our resources, it effectively provides us with another legal team”.
By working in tandem, the humans and the machines benefit from each other’s unique capabilities.
And of course, machines are not infallible. Amazon reportedly used AI-based recruiting software to help review CVs and make recommendations.
But the software favoured male applicants because it was based on historic CVs submitted to Amazon when many more male candidates were hired.
Applications that contained certain triggers – the word ‘women’ or attendance at a girl’s school – were reportedly downgraded by the system.
Perhaps rather than job losses, a primary benefit for humans from increased use of AI will be a happier workforce with greater job satisfaction.
Let’s face it, there are some tasks that we might prefer not to do – large data examinations or extensive discovery reviews.
With machines doing the heavy lifting of the mundane and monotonous tasks, intelligent and educated people can concentrate on more fulfilling work where they really add value.
So document automation tools can be used for the process driven drafting in a sale agreement (changing ‘Seller’ to ‘Sellers’, ‘Company’ to ‘Group’, incorporating standard provisions for a split exchange and completion, etc.) while the lawyer gets on with drafting the bespoke consideration provisions, such as tailored completion accounts and a complicated earn-out mechanism.
Reducing the monotony could also reduce the risk of a bored or tired lawyer making a mistake.
With the rapid changes taking place in the legal sector, firms are being forced to adapt and innovate to succeed.
But it’s important to remember that while machines may do the work, it will still be the humans who bear the responsibility.
As Julia Hoggett, FCA Director of Market Oversight, put it: “the FCA cannot prosecute a computer, but we can seek to prosecute the people who provided the governance over that computer”.
Sophie Brookes is a partner at Gateley gateleyplc.com