.png)
.png)
Creativity and innovation are at the core of human existence. Throughout history, we have invented technologies that have completely transformed our world. From the plow to the wheel to the combustion engine; time and time again humanity has found itself on new frontiers of science, and society has adapted around these advancements. Artificial intelligence (AI) is the invention that will cause the biggest changes to our 21st-century lives, as it poses to automate jobs and industries. But what exactly is AI, and should we be rushing to embrace it?
The term ‘Artificial Intelligence’ was coined in the 1950s but it didn’t enter our vernacular until much more recently. When we talk about AI, we’re referring to a non-human machine (that’s the ‘artificial’ part) that thinks as a human would (the ‘intelligence’ part). An image of humanlike robots living among us, ending civilization as we know it often springs to mind when people visualize AI, whereas, in reality, we needn't be concerned. AI will likely never manifest to end humanity, however, for us lawyers it will still significantly disrupt our working lives. Here’s how.
One of the greatest shifts to the legal profession our generation has seen is the growth in demand for legal services from large corporations, and the resulting influx of lawyers into in-house roles to support that cost-effectively. In the years leading up to the global financial crisis (GFC) of 2008, Am Law 200 firms went on a hiring spree. At the time, legal software was rudimentary and so the provision of legal services was very manual. The growing economy meant that large corporations were consuming more and more legal advice, and the best way to win and profit from that work was to amass many top graduates from elite schools to complete the work.
When the GFC hit and the economy began to shrink, the demand for legal services was still there, but the budget was not. Corporations were no longer willing to pay such high rates for junior lawyers and instead began to grow their in-house teams and reduce their reliance on law firms. Large corporations today typically have teams of technology-savvy and commercially minded in-house lawyers tasked with managing their external legal spending. A 2020 Thomson Reuters LDO survey into the demand for legal services during the pandemic shows that controlling those costs was identified as a high priority for the vast majority of legal departments.
Not all legal work is well suited to outsourcing, and corporations with a large in-house team can be selective about the types of tasks they choose to send externally. This work often consists of a high volume of standard ‘run the company’ tasks that the team cannot scale to absorb themselves, alongside a smaller portion of bespoke ‘bet the company’ work that requires very specific legal expertise.
The Thomson Reuters LDO survey demonstrates how the demand for legal services from within corporations does not always align with what is passed on to law firms. Law firm respondents reported a drop in demand for legal services during 2020, but that drop was not shared by clients who were inundated with work that “required an in-depth knowledge of their businesses and a very quick turnaround”, making it unsuitable for outsourcing.
For the legal work that is still being outsourced, while a traditional law firm may be the client-facing provider, in reality, the high volume and low complexity nature of the work, coupled with the very competitive market for providing legal services, means that the firm is probably not delivering those services in an entirely manual way anymore. How is this possible, when just a few years ago we relied so heavily on human power?
The most simple explanation is that advances in computing power and the cloud have made it feasible for software to use AI to complete some of the steps that a junior lawyer would typically take to complete a task manually.
Once cloud-hosted servers became readily available at an affordable cost, software provided through web applications hosted on the cloud began to be accepted as an alternative to locally installed software. This allowed legal software providers to take advantage of the availability of much greater power to create systems that could handle the complex operations needed to incorporate AI into software.
AI will most notably impact our daily lawyering by making us more efficient. For example, a lawyer can use a contract review system to analyze a contract in just a few minutes when they’d previously have taken much longer to print the document, highlight the key provisions, and then type their comments back into a summary. Time savings are priceless to an individual lawyer and legal department that are inundated with work, as such the potential for increased efficiency is very positive.
These AI-powered systems cannot complete legal tasks entirely automatically but instead, focus on using a lot of computing power to automate the most mundane parts of the task. A good example is finding a change of control provision in a contract. A human who was going about this task would need to know what a typical change of control provision looks like, what it is that makes it a change of control provision rather than something else, and then methodologically read through each page to find a provision that matches.
A system completing this task takes a similar approach. First, it collects many examples of the change of control provision that it uses to determine the common characteristics of the provision. It then stores this set of criteria as an algorithm. When it needs to review a contract, it uses the algorithm to find the patterns that are common to the provision and systematically applies this algorithm to each paragraph in the contract until it finds the provision which matches.
In this scenario, using AI system to find provisions in contracts rather than doing so ourselves has disrupted our lives as lawyers, and when applied en masse will disrupt the legal industry as a whole. From an in-house consumer’s point of view, we still have a lawyer providing legal services but we have significantly reduced our demand for that lawyer’s time by redesigning the process, albeit relying heavily on a system to do so.
A 2018 LawGeex experiment pitted their AI-powered contract review system against lawyers reviewing the same agreements. In that study, the system completed the review of five contracts in 26 seconds with an accuracy of 94 percent, whilst the lawyers took 92 minutes with an average of 85 percent accuracy. On the first impression, this study seems to paint a rather grim picture for lawyers, which is unfortunate because it is not reflective of how AI systems work in practice.
While we can expect systems to expedite certain parts of the process, infusing them with enough intelligence to make lawyers entirely redundant is not realistic, but more importantly, it is not desirable. In short, the most valuable thing to the client is not an entirely automated solution or an entirely human solution, it’s an educated and intelligent lawyer augmenting their processes with a smart system to save time when it makes sense to do so.
The key areas in which AI will be most visible to lawyers are the tasks that are low in complexity and high in volume. Contract review, e-discovery, and legal research are good examples of tasks that have the potential to be completed much more efficiently with intelligent systems. The review of executed contracts necessary for M&A due diligence has already been made much more efficient by contract review systems that can analyze huge quantities of contracts at speed. Analysis of pre-execution contracts is also an area to watch. According to Forbes, “companies are developing AI systems that can automatically ingest proposed contracts, analyze them in full using natural language processing (NLP) technology, and determine which portions of the contract are acceptable and which are problematic.”
A much more contentious use case is in the potential use of systems to predict outcomes in court proceedings. By analyzing data from past cases, systems can theoretically utilize supervised machine learning to forecast future results. Whether we’d want them to is a different question.
For those of us who work in-house, we are perfectly placed to take advantage of the benefits of intelligent systems, both in-house and in the work that we outsource to law firms. In-house legal teams are not selling their services back to the business by the hour and so any system that can help them work more efficiently is an easy sell. Asking external counsel to implement the same systems to reduce the cost of the services they’re providing is a more difficult proposition. To get the most from law firms we need to empathize with the change we’re asking from them and hold ourselves up to those same standards we’re imposing.
