We have all seen how the interest in generative AI has taken the world by storm over the past 12 months. However, we are only just starting to see how and where these tools can be applied to legal practice. Legal has arguably tinkered around the edges with some automation tools for years, but generative AI (or Chat GPT as most still refer to it 😊) has brought the AI ‘conversation’ into the mainstream. [note my use of conversation, rather than more widespread use cases].

However, not so when it comes to eDiscovery.

eDiscovery has deployed AI in the form of Technology Assisted Review (TAR) or Predictive Coding for many years, even if the adoption was not as widespread as one would have hoped.

Why the discovery process needs this technology?

The existing discovery model is largely based on traditional practices that are no longer sustainable with the proliferation of information we have today. The challenges and costs can easily spiral out of control, as we have no option but to look at ways to simplify the process and not just do things as we always have. The volumes and sources are not decreasing anytime soon.

That is all before even contemplating the elephant in the room, that the discovery process has for far too long been a good money earner for law firms, especially to fully utilise junior resource. I do not need to get into the merits of this (I would love to…😊), which should be evident for all, but likewise the client doesn’t want this and lawyers do not want to be using their legal expertise for page turning documents.

The objective of the discovery process should be to get only what you need and do so in a way that is quick and cost effective.

Too often what held TAR back, was the fight against the perception of the gold standard traditional eyes on every document approach that has simply led to so much needless cost over the years. The skill is to come up with methods and leverage the use of technology to get rid of what you don’t need so that you can devote your energies at only looking at what matters most.

Any option that helps us get to the most important information quicker and cheaper, whilst helping to isolate irrelevant material, cannot be a bad thing !

This is where TAR can be a solution!

How does the AI works?

Very simply, Technology Assisted Review (TAR) works like this –

Subject matter experts (i.e. lawyers) reviewing and making calls on relevance, with the computer using algorithms to learn these calls and applying the calls to a wider set of documents. It is an iterative process that continues to learn as more information is looked at.

Used correctly (again emphasising correctly…), the software does not make the final decisions on the documents but prioritises what the lawyer should look at further. Like Netflix and Google, it is making recommendations based on your prior choices – in this case how you have interpreted the relevancy of documents.

The technology is not new as it has now been around for the best part of 10-15 years, helping to combat the challenges of the discovery process.

We even put it in the NZ discovery rules !

Even when we put together the New Zealand High Court Rules (HCRs) back in 2010 (and introduced in 2012), we set the framework for the acceptance and use of AI technologies like TAR, and encouraged their consideration to help improve the efficiency of the discovery process.

Often what restricted TAR adoption was the work required upfront with training and protocols, whilst also justifying its use. The shift to more continuous active learning, whereby the technology continues to learn as the review progresses has helped simplify the use and adoption of TAR.

As TAR is always running in the background, you can choose if or when to use it.

I have found this beneficial when at the outset, as the legal team may not be certain about deploying TAR, but know that they can utilise it later if they choose. Inevitably I have found on large matters, the onerous nature of the exercise, may only be realised once the review is well underway. The advantage is they can quickly deploy TAR, which has been working in the background, fully using the existing work of their review – nothing lost, plenty gained !

The lawyer drives the process

The key to the success of TAR is that it puts the lawyer at the forefront of the process, relying on the human judgement and critical thinking of the lawyer. The technology is helping the lawyer make better decisions, and much earlier in the process with quicker access to the key information – all without having to spend considerable time and money.

The technology learns and makes decisions based on the knowledge of your subject matter expert. Of course as with any AI, ‘garbage in garbage out’ applies – as long as the person(s) making the calls have the right expertise and knowledge of the matter, or else…

Anyone that has undertaken a large scale document review with multiple reviewers will know too well the issues with inconsistencies of the calls. Too often this perceived gold standard is so far from being anywhere close, and that is before we even consider the costs of such an approach.

A great enabler in eDiscovery

One of the great advantages of technologies like TAR is it provides a great enabler for those that may not have as many resources at their disposal. They can simply use the expertise of their one or two lawyers to compete with much larger firms, and in doing so handle much larger document volume matters. From my experience most of the clients I have worked over many years utilising TAR have been smaller firms, that are usually up against much larger law firms.

I know from a personal perspective, we have had some fantastic success confronting the bigger players !

The success of AI in the discovery process can enable a better outcome, without the lawyer being weighed down by the time-consuming discovery process. If you are not at least considering the use of AI technology like TAR, you or your clients are missing out on tools that can help move through the discovery process quicker, cheaper and more accurately.

AI in eDiscovery will evolve further

AI in eDiscovery is only getting started. We all know how much technology can change and quickly.

Let’s hope with this interest in AI now hitting the mainstream gives a renewed adoption to the likes of TAR, or any tools that can help improve the discovery process. Sometimes we simply lose sight of the tools that have been available to us for many years.

It will be interesting to watch how generative AI may impact eDiscovery further, as we are now just starting to see how eDiscovery solutions are to apply generative AI capabilities within their platforms – of course far beyond just TAR.

Definitely exciting times to come with eDiscovery in how generative AI may be deployed further !