The AI and Investigations Paradigm: Faster, More Precise, Defensible
The average e-discovery matter today involves more than 6.5 million pages of documents. In complex, multi-national matters and high-stakes internal investigations, data volumes are often much higher. Outdated manual processes and reluctance to leverage advanced tools to find key information from within large, diverse data sets are resulting in excessive time, cost and complexity in bringing matters to a swift and defensible conclusion.
Amid the ramping demands of investigations today and the expectation for legal teams to reduce costs, many lawyers are still reluctant to use AI-enabled analytics platforms and workflows. FTI Technology and Relativity’s 2022 General Counsel Report, a survey of in-house counsel, reported that 73% are not using Artificial Intelligence (AI) for any function and only 33% said they agree attorneys currently have adequate technological knowledge and capabilities. Another report from CLOC found that only 12% of corporate legal teams report using advanced analytics and technology-assisted review (TAR).
One common refrain is that AI is powerful beyond limits — which leads to intimidation and worry that the results may be inaccurate or difficult to defend in a dispute. While the tools are indeed powerful, they are not a black box. The technology can be understood and validated using data science and statistical methods. When applied by experts using proven workflows, the results are both highly accurate and defensible. While it may seem like a safe play to delay adoption until AI and other advanced tools reach wider acceptance or lawyers gain more technical proficiency, the reality is that the next generation of e-discovery is arriving. Ready or not, matters are quickly and consistently becoming more complex.
With growing data volumes and emerging data sources introducing new challenges in litigation, investigations and other legal and regulatory matters, the use of AI will be essential to efficiently prioritize documents and quickly surface key facts. AI tools can also scale processing for large quantities of data, cull irrelevant documents and enable review quality control. AI-powered workflows within leading platforms like Reveal, Relativity and Nuix Discover consistently outperform manual approaches in terms of time savings, reducing document sets and streamlining costs.
Implementation of AI—for investigations, in which the team has an idea of what they are looking for, but need AI to quickly surface patterns or behaviors of interest across a large dataset, compliance monitoring, due diligence and other matters—is often the most strategic and impactful way for legal teams to meet the demands they face. Top advantages and opportunities AI tools bring to investigations use cases include:
- Allows investigators and document reviewers to begin examining the data for patterns and key insights right away, as no up-front model training is needed.
- Can efficiently capture relevant documents without the need for search terms.
- Handles rolling data loads seamlessly, as the addition of new document sets or inclusion of new custodians does not require any new model training or sampling.
- Ability to find facts and build an understanding of what happened even in “low richness” scenarios where the team is not guided by a defined set of keywords or date ranges.
- Bias and sentiment analysis to determine whether communications contain positive, negative or neutral expressions and/or repeat instances of certain sentiments.
- Clustering of similar terms to make connections between documents containing similar relevant or notable content.
- Social network displays to illustrate the web of communication between key custodians of interest or certain activities of interest, so investigators can easily find abnormalities and red flags.
- Ongoing learning as investigators and reviewers make decisions about certain documents—every human decision is captured and supports further refinement of the tool’s findings.
- Portable models that take previous learning or prebuilt algorithms and apply them to a new matter, use them to improve another model or set them as the foundation for a model being built from scratch, without needing to do any new training.
In one recent matter, our team scaled up a multi-national investigation spanning multiple countries, more than 300 custodians, roughly 70 million documents and a time period of more than 17 years. The use of AI was critical to scale up the matter and allow the investigatory team to begin examining the facts without the typical delays associated with processing large volumes, training analytics models and testing samples. If traditional search term based investigation approaches would have been followed, that would result in more than 3 million hit documents for review. By scaling up with AI, the team has identified hot and relevant documents without search terms. Among roughly 300K documents identified as very relevant to the case, approximately 30% of them without search term hits have been captured in a more efficient way thanks to the well trained models.
AI can offer tangible benefits in disputes and investigations. The advancements we’re seeing in the available technology have the potential to significantly drive down the time, cost and complexities of dealing with large data volumes and the technical diversity of data in the wild. With the severity urgency of investigations rising worldwide, advanced technology applied alongside experienced practitioners will be increasingly important to containing the time, cost and risk of these high-stakes matters.
The views expressed herein are those of the author(s) and not necessarily the views of FTI Consulting, its management, its subsidiaries, its affiliates, or its other professionals.