When asked what the pros and cons of using artificial intelligence (AI) for the disclosure process in litigation are, ChatGPT produced four positively-skewed and four negatively-skewed answers. While some of the points listed by ChatGPT were similar to those raised by the litigator attendees at a roundtable dinner hosted by The Lawyer in association with legal tech firm DISCO, the AI did not debate them quite as passionately as the human lawyers did.
One of the most significant changes litigators said they have witnessed over the past 30 years is the immense influx of documents to review as part of the disclosure process. Gone are the days of trial bundles that contained only 100 documents – litigators are now dealing with them in the millions. This is why the use of AI-powered technology for e-disclosure has been so welcomed, as it has saved legal professionals vast amounts of time, reducing costs as a result.
But with competition between firms as fierce as ever, litigators must demonstrate to clients that their understanding and use of technology to innovate is head and shoulders above the rest – especially when it comes to the disclosure process. With that in mind, here are the key takeaways from the roundtable discussion, in which litigators from top City firms to specialist litigation boutiques attended.
Training the AI
One of the cons listed by ChatGPT on using AI for the disclosure process in litigation was that AI can reflect the biases and assumptions of the data it is trained on. If the training data is biased or incomplete, then the results produced by AI-powered tools may also be biased or incomplete. Training the AI for e-disclosure was a key issue among some litigators when using Trial Assisted Review (TAR), with one partner commenting that it sometimes feels like using an “evolutionary rather than revolutionary model”.
Another attendee articulated that training a system can be very difficult when you have many issues for disclosure: “Say you have 20 issues that people are looking at. It’s very difficult for one person or a small team of people to compartmentalise their minds when they’re clicking through to train a system properly. This is a big missing piece, and I would be less comfortable doing the selling pitch of TAR to clients or an opponent until I was comfortable that it worked in that context.”
Another critical issue was the feeling that both litigators and their opponents needed to be fully aligned in using it: “It does work, and we should be selling it,” said one attendee. Another responded: “If we are going to make better headway with it, we’re going to need more transparency and better common standards.”
Getting clients to trust in tech
One partner referenced that the first female judge in the Commercial Court, Elizabeth Gloster, said a decade ago that it was only by embracing technology that lawyers would stay competitive in the disputes area in London. But when embracing technology, lawyers have the additional hurdle of getting their clients on board, and the topic of client trust was still a split issue around the table.
Some expressed that client trust was still lacking in how the search is run and that there was still some way to go to help lawyers explain the benefits and methodologies. Without proper explanations and justifications, the costs could go up rather than down. As one lawyer commented: “If the client doesn’t trust it, or the lawyer doesn’t, then you’re duplicating time, and costs go up.”
Another described clients as falling into two categories: some clients can appear more reticent as they don’t have much experience using the tech and therefore are not as interested. Other clients, typically those with sophisticated in-house legal functions, are miles ahead and already have their chosen vendors and systems in place.
Philip Demetriou, DISCO’s senior director of international sales, thought they would see more clients driving law firms to start leveraging AI: “We are seeing a lot of corporate clients clamouring for the application of technology and AI within the field of litigation and arbitration.”
The ‘smoking gun’
One positive not mentioned by ChatGPT, which is perhaps because it does not have the lived experience of being a litigator practising in England, is that the disclosure process in litigation is what makes England stand out as a legal hub to clients.
One partner said: “Before accepting the different tech systems used for disclosure, the client has to buy into the value of disclosure in the first place. We now have competitors in the Netherlands, France and Germany. People are saying litigation is going away from the UK and off to Europe. But there’s one thing they don’t have: disclosure. If the client litigates in the Netherlands, they won’t find that smoking gun evidence among the five million documents at all.”
Another partner reiterated this point, emphasising that among the five million documents, a handful of those documents or a critical email, might be “the one thing that turns the trial.”
Sponsor’s comment: Richard English, director of review operations at DISCO EMEA
As a provider of AI-driven legal solutions, we at DISCO take an evangelical stance on the role of technology in law. There seems to be a consensus that 2023 will mark the advent of broad adoption of AI-enabled tools and services across many industries, and Legal is no exception: take for example the booming interest in TAR2.0 workflows in edisclosure.
The question we wanted to understand by hosting this event in partnership with The Lawyer, was: What do seasoned litigators think of technology adoption in legal practice? Are large language models and generative AI seen as the “bogey man” who will encroach on traditional litigation work? Or do litigators share our own stance?
It was therefore refreshing to be reminded that lawyers have been adopting new technologies along with every other industry, and that bespoke “legal tech” is a topic of high interest. Discussions flowed throughout the evening as anecdotes were shared of experiences using technology in practice. It was acknowledged that lawyers are naturally circumspect when it comes to risk, and this extends to their own practices. However, provided these risks were mitigated against, then tech-enablers will be welcomed.
Richard English is Director of Review Operations at DISCO EMEA. With over 13 years experience in the field of compliance, discovery and document review, coupled with a background in the legal sector as a South African attorney, Richard is a veteran in the field with extensive insight into how to run and manage successful and significant review projects.
