Making sense of AI in eDiscovery: How new technologies are streamlining workflows
Key benefits of AI in legal workflows
Increasingly, corporations and law firms face the formidable challenge of managing legal and compliance risks amidst the proliferation of complex data. The pressure to do more with less pushes teams to seek innovative solutions. Emerging AI technologies offer a promising avenue, potentially transforming how legal and compliance tasks are approached and managed.
For leaders of these organizations, however, knowing how to leverage these technologies can be overwhelming. By better understanding and leveraging AI and other emerging technologies, firms can streamline processes, enhance accuracy and ultimately reduce the burden of handling vast amounts of intricate data.
From TAR to Gen AI: A journey of innovation
The field of eDiscovery has undergone a significant transformation over the past decade, largely driven by advances in AI. Each evolution and iteration of this technology has introduced new capabilities and efficiencies.
Technology-Assisted Review, or TAR, refers to using machine learning algorithms to help identify relevant documents during the eDiscovery process. This is an example of Discriminative AI, where the model classifies documents (e.g., responsive vs. non-responsive) but does not provide further analysis.
The first TAR matter in the U.S. is generally recognized as case Da Silva Moore v. Publicis Groupe, in 2012. Magistrate Judge Andrew Peck of the Southern District of New York issued an opinion that endorsed predictive coding (a type of TAR) for the first time in a U.S. court.
Similarly, Pyrrho Investments Ltd v. MWB Property Ltd in 2016 marked the first major judicial endorsement of predictive coding in the U.K., significantly impacting the acceptance and use of TAR in European legal systems. The decision set a precedent, encouraging wider use of TAR in electronic discovery in Europe, just as the Da Silva Moore case did in the U.S.
TAR offers a variety of key benefits, including predictive coding, which enables the system to learn from expert reviewers and distinguish relevant from irrelevant documents. This increases efficiency by reducing the volume of documents needing manual review, saving time and costs. TAR has gained acceptance in courts through several landmark cases affirming its reliability.
TAR2, the next iteration of the technology, uses continuous active learning (CAL) instead of a single training round, enhancing accuracy over time. This iterative process allows for ongoing updates, making TAR 2.0 more adaptable to the dynamic nature of legal cases.
Conduent, in collaboration with its clients and their law firms, has achieved significant success with TAR2, using our proprietary Viewpoint eDiscovery Solution in conjunction with Relativity. Initially, a more cautious approach was adopted, with a full review of all search hits. We employed TAR2 to prioritize the review of responsive documents, consistently identifying 90% of these documents within the first 10% of the review process. In our case, as confidence in TAR2 grew, clients experienced up to 90% savings in review costs. This was due to the early identification of responsive documents and subsequent validation through allusion testing, eliminating the need to review every document.
The future of AI
Generative AI represents the next frontier in AI technology. For eDiscovery organizations, Gen-AI leverages these capabilities to improve document review and analysis. Key features and impacts include:
- Natural language understanding: Gen-AI models like GPT-4 and LLAMA2 excel in understanding, enabling more intuitive and effective document analysis.
- Automated summarization: These models can automatically generate summaries of large document sets, highlighting key information and trends.
- Knowledge retrieval: Gen-AI allows users to question their data using natural language, eliminating the need for unreliable search terms and complex Boolean searches.
- Removing busy work: Gen-AI supports mundane processes like creating privilege logs, redactions, and data normalization, tasks traditionally handled manually or with partial automation.
Partnerships matter
When planning AI-driven solutions for your organization, it’s critical to choose a leader experienced in serving firms and organizations of all sizes. Conduent has been a leader in AI-driven eDiscovery for over a decade. Our mission is to deliver industry-specific AI solutions within a secure environment, ensuring high value at a reasonable cost. We have achieved this through strategic partnerships with leading law firms and corporations. The evolution of AI technologies presents opportunities and risks, which can be effectively managed through proper planning and collaboration.
Ready to harness the power or AI to optimize contract reviews for your organization? Visit us online to learn more, or reach out to an expert now at https://www.conduent.com/business-operations-solutions/legal-compliance/.