Recent surveys of the legal market reveal there is a marked disparity between the high interest in AI tools and technology and the relatively low current level of actual deployment. This gap is evident across the market, among lawyers in private practice and corporate law departments, with the notable exception of the largest law firms, where there has already been a significant adoption of various AI tools for the purposes of due diligence, contract review and AI assisted e-discovery. But on the whole, adoption continues to lag, with 54% of all law firms participating in the most recent ILTA survey indicating that they are not presently pursuing AI options and a much lower 12% of corporate law departments indicating that they currently use AI tools and processes, according to the latest CLOC survey.
So why is AI adoption still lagging in the legal market? Of course, we can always blame it on the legal market’s traditional aversion to technology but frankly that is wearing thin as an explanation. It’s 2020 and even the Luddites among us are chatting on their iPhones. Technology has pretty much gained ascendancy over every aspect of our working and personal lives. So it’s reasonable to infer that there must be some factors that are very specific to AI technology that are responsible for slowing down its adoption in the legal market.
As it turns out, one of the major obstacles to adoption of AI/ML tools is just how labor intensive these projects can be, at least at the outset, due to the need to undertake adequate system training. AI/ML systems rarely work smoothly out of the box, especially not when it comes to applications in the legal market, which entail processing large amounts of unstructured data. In fact, training AI/ML systems to parse contracts and other legal documentation usually requires the data to be properly cleaned, tagged and annotated in order to make it readily machine readable. The effort and complexity of the task is far from trivial, as it depends on human review teams that are not only facile working with the applicable AI tools but can also properly tag and code the contract data. The effort may be especially onerous for corporate law departments because they will typically have more extensive training requirements, in order to adapt the AI system to their very specific environment.
So there’s something of a Catch 22 at work if a corporate law department hopes to launch a successful AI initiative. If you’re looking to benefit from the cost and labor saving potential of AI and machine learning, you are going to need access to highly skilled and trained staff in order to get your AI initiative off the ground in the first place. Understaffed as most law departments are these days, finding the right outsourcing partner may very well turn out to be a crucial factor in determining your AI project’s ultimate success.
We had a chance to speak with Adi Mirza, the CEO of Cenza, which is one of the legal outsourcing companies that has emerged as a key resource in the AI/ML marketplace, providing these system training services on an outsourced basis. Cenza’s clients include AI software developers and service providers, as well as corporations and law firms looking to implement the technology.
“People who don’t have first-hand experience working with AI/ML tools are prone to a basic misconception about the way they work,”
as Mirza explains it.
“You don’t simply flip a switch and presto the system performs magically. AI/ML systems need to be trained and fine-tuned in order to perform at the high level of accuracy that is required for most business and legal applications. This means that system training and very often ongoing human review of system output is absolutely necessary to insure a successful AI implementation.”
The evolving role that companies like Cenza are playing in the market highlights the dramatic impact that AI technology is already having on the legal industry’s supply chain.
“In the last few years, there has been a significant transformation in the sort of work we do,”
according to Mirza.
“A growing percentage of our clients come to us now looking to leverage AI tools and technology in the way they handle contract management and document review projects. As a result, our role is shifting. Outsourcing services are no longer driven by the simple arbitrage of low-cost wages through offshoring. We’ve had to develop expertise in working with all the major AI systems in the legal market, such as Kira, Leverton and Ironclad , so that our review teams can now work seamlessly with whatever technology the client is using. The big pay off in efficiency comes through the integration of machine learning and human review teams.”
Every new technology comes along with its own unique jargon, acronyms and terminology, and AI/ML is certainly no exception. Human-in-the loop is one of the phrases you hear frequently when you talk with Adi Mirza and others who are on the leading edge of implementing these new AI assisted workflows. The Cenza review teams that tag, code and annotate contracts are the humans-in-the-loop that are necessary in order to enable system training and the growing power and sophistication of AI/ML tools and technology.
This is the new reality of the workplace. Over time, legal service delivery will be reshaped from end to end, as AI tools are deployed to handle an increasing number of more sophisticated tasks and operations. As that happens, it becomes increasingly important to be able to adapt and upgrade your skills, the way the human review teams at Cenza have done, so you don’t get run over or made obsolete by the new modes of legal service delivery. One way or another, we all need to be ready to function more effectively as humans-in-the-loop, ready and able to adapt our human intelligence so we can take best advantage of the artificial tools and technology that will be readily at our disposal .