How AI Is Changing Pretrial Discovery

Pretrial discovery often accounts for most of the costs of civil litigation. Even before the rise of artificial intelligence, the discovery process evolved substantially over the past 20 years. For example, electronically stored information (ESI) — such as electronic documents, emails, databases, and metadata — has become the primary source of discoverable material. The use of e-discovery platforms has streamlined the management and review of documents.

The rapid development and launch of AI tools for legal professionals is spurring additional changes to pretrial discovery. As AI tools reduce the costs and/or increase the accuracy of document review and the broader discovery process, law firms, in-house legal teams, and legal software providers are expected to adopt these tools more widely.

Enhancing Document Review Workflows With AI Tools

The most immediate impact of artificial intelligence on pretrial discovery has been the incorporation of AI models and features into existing e-discovery platforms, as well as the launch of new, AI-first discovery automation platforms.

The pattern-recognition capabilities of generative AI and large language models can sometimes uncover subtle inconsistencies and other nuances in documents. Machine learning algorithms have demonstrated an ability to be trained to identify documents by privilege, relevance, and responsiveness for further review by document review attorneys.

Like many other software providers across a wide variety of industries, established e-discovery providers have been quick to incorporate AI capabilities into their platforms and take steps to build customers’ confidence that they are staying ahead of the curve. New providers touting impressive AI features are also taking advantage of the moment to seek greater market share.

Get the Right Legal Help for E-Discovery

E-discovery procedures are complex. The Super Lawyers directory contains the best attorneys in the field who can assist you with this process.

Find a lawyer today

Using Legal AI Tools Responsibly

In-house legal teams and law firms that want to make responsible use of new AI tools to enhance their document review processes will need to

• Critically evaluate the claims of providers
• Test AI system capabilities to make informed decisions about whether and how to use them
• Ensure that any AI systems they use have adequate security safeguards
• Update their workflows accordingly

One particular area of risk is that AI systems tend to be responsive to subtle nuances in the phrasing of prompts. A carelessly assembled prompt may cause an AI model that is otherwise fit for the purpose for which it is being used to overlook relevant information and produce lower-quality results.

Law firms and legal departments are finding that it is particularly important to incorporate prompt engineering and review steps into their AI document review workflows.

Using AI to Assist in Deposition Practice

Another area where AI technology has impacted pretrial practice is depositions, particularly the preparation stage.

Attorneys have found the pattern-recognition capabilities of generative AI tools useful in identifying key issues raised by document datasets, generating questions, preparing outlines, and anticipating objections.

AI systems have also shown some promise in deposition review. In reviewing large volumes of deposition transcripts, AI models have the ability to flag both direct contradictions and more subtle potential inconsistencies in a witness’s testimony or between the testimony of multiple witnesses. There is also some potential for AI tools that are given access to deposition video recordings to help attorneys identify nonverbal cues that may help impeach a witness.

New Risks and Complications Arising From the Growth of AI

The use of artificial intelligence, both in pretrial discovery practice and in broader business communications and documentation, does generate some new risks that attorneys need to attend to.

Attorneys Are Responsible for Work Product, Even When Delegated

First, as always, attorneys are professionally responsible for their work product, and that responsibility cannot be delegated to an AI service provider, much less an impersonal algorithm.

Anyone who uses AI to assist in legal practice must gain an understanding of how these tools work, document the steps they take, and independently verify the validity of any AI outputs. Attorneys may also have an ethical responsibility to disclose their use of AI to clients.

Attorneys Have a Duty To Check AI-Generated Content

More broadly, the widespread use of new and often poorly understood AI tools in daily business communications, presentations, and workflows introduces some complications into the document review process.

Documents that appear, on the surface, to have been prepared by a human may have actually been generated by an algorithm, and they may contain inaccuracies and misrepresentations, sending document review attorneys down false lines of inquiry.

Additionally, the use of AI systems to review documents that may have, in turn, been generated by AI could produce unexpected results. These complications will undoubtedly require further refinement of document review and pretrial discovery workflows over time.

Staying Abreast of Legal AI Developments

While AI has already had a substantial impact on pretrial practice, much remains to be seen about whether providers can deliver to anticipated future capabilities and whether ethical considerations will require courts and bar associations to adopt new rules governing its use.

Attorneys owe it to themselves and their clients to stay abreast of these developments.