Discovery disrupted: will generative AI prompt more civil lawsuits to go to trial?

September 2026  |  SPOTLIGHT | LITIGATION & DISPUTE RESOLUTION

Financier Worldwide Magazine

September 2026 Issue

The number of civil lawsuits that go to trial has been on a steady decline for decades. Although several hundred thousand federal lawsuits are filed annually in the US, few proceed to trial. Data from the Administrative Office of the US Courts reveals that for the year ending 31 December 2025 only 0.6 percent of terminated federal civil cases were resolved during or after trial.

The scarcity of civil trials is often attributed to the cost of the discovery phase of litigation. The high price tag for electronic document review (e-discovery) can make it financially prohibitive for a corporate defendant to try a case, even one that it might be able to win at trial. The economics of discovery frequently make the business case for settlement more compelling than the prospect of a victory at trial.

While technology assisted review (TAR) has increased proportionality in e-discovery, disproved the notion that manual review of all documents is the gold standard, and reduced document review costs, it has been unable to move the needle enough to prevent e-discovery costs from being obstacles standing between corporate defendants and civil trials. Artificial intelligence (AI), however, has already had a dramatic impact on the practice of law and may have the potential to transform discovery to the point that it ceases to be a driving factor in the decision whether to try a case or settle it. Whether AI can achieve this result remains to be seen, but if it shows signs of bringing about that transformation, should businesses, lawyers and courts prepare themselves for an increase in the number of civil cases proceeding to trial?

The economics of discovery

The purpose of pre-trial discovery in civil cases is to give the parties the opportunity to gather evidence, develop a factual record and prevent surprises at trial. The growth in the volume of email and electronic business records maintained by companies along with the increase in discoverable cell phone and social media data created by individual plaintiffs, has made the review of electronic documents an expensive, time consuming and critical part of pre-trial discovery.

In a 2012 study, the RAND Corporation for Civil Justice found that document review accounted for 73 percent of total e-discovery spending. A study by Complex Discovery in 2024, following the adoption of TAR by law firms and corporate legal departments, showed that percentage at 64 percent. According to industry projections, in 2029, review of electronic documents will continue to be the most resource intensive stage of e-discovery but will account for only 52 percent of total e-discovery spend, presumably due to the expected adoption of AI-driven e-discovery tools.

From document review to litigation intelligence

Generative AI (genAI) is advancing e-discovery beyond TAR, providing litigation teams with additional efficiency-enhancing capabilities, including the ability to ask natural language questions of document collections and generate explanations for coding decisions. It is being used for first pass document review for relevance, confidentiality and privilege, creation of privilege logs, quality control and early case assessment. GenAI powered discovery workflows are reducing the cost of summarising documents and deposition transcripts, developing chronologies and case themes, identifying and contextualising hot documents, surfacing contradictions between witnesses and data, and compiling witness files and deposition exhibits.

Governing AI before it governs your case

GenAI tools’ potential to slash e-discovery costs may be tempered by the ongoing need for legal teams to be involved in these AI-assisted workflows. While cost savings may be achieved through smaller legal teams spending fewer hours on document review, the role of lawyers will arguably be amplified as they take on responsibility for model training, coding, validation and oversight. Recent cases highlight the importance of lawyers exercising technological competence and advocacy skills when AI is used in discovery.

In White v. Walmart, Inc., the plaintiff’s counsel challenged the sufficiency of the defendant’s interrogatory responses based solely on a list of alleged deficiencies he generated using AI. When questioned by the court, the plaintiff’s counsel admitted he had created the list by entering the defendant’s discovery responses into an AI programme and asking the AI to determine which responses were insufficient. The plaintiff’s counsel then copied and pasted the output generated by the AI into an email that he sent to the defendant’s counsel and the court.

The court found the plaintiff’s counsel’s exclusive reliance on AI to be improper, stating that “[a]ttorneys and litigants must exercise independent judgment and oversight in discovery-related matters, even when using AI tools”. Acknowledging that the “use of AI platforms is not itself problematic”, the court characterised the plaintiff’s counsel’s conduct as taking “a perilous shortcut around his responsibilities as a trained legal professional”.

In another recent case, Conservation Law Foundation, Inc. v. Shell Oil Co., a magistrate judge ordered the plaintiff to produce the AI prompts its expert witness and her team used in preparing her report. The expert had used a commercially available genAI tool to filter and identify potentially relevant documents from the defendant’s document production. The defendant moved to compel production of the prompts the expert used to conduct her AI analysis.

The court rejected the plaintiff’s argument that the AI prompts used by the expert were outside the scope of discovery set out in Federal Rule of Civil Procedure 26(b). The magistrate reasoned that, under the facts of this case, “the process by which the expert had culled down the defendant’s document production into a subset to be worked with is an aspect of [her] methodology” and, as an expert’s methodology is fair ground for discovery, the AI prompts must be produced. Believed to be the first federal court ruling ordering the production of an expert’s AI prompts, the order has been stayed by the district court.

AI will change discovery practice, but not necessarily trial statistics

AI may reduce one of the largest barriers to trial, but others remain. Although AI will continue to streamline e-discovery workflows, enhance the capabilities of legal teams and reduce document review costs, it is unlikely to lead to a dramatic increase in the number of federal civil cases going to trial in the foreseeable future.

The potential financial and operational consequences of an unfavourable outcome at trial and the risk of reputational harm from a public trial will continue to make corporate defendants hesitant to take cases to trial. AI may be able to provide insight into how a judge is likely to rule at a bench trial after reviewing the judge’s prior decisions and controlling legal precedent, but it cannot reliably predict whether a key witness will perform poorly on the witness stand at trial or whether a judge will find flaws in an expert’s opinion. The inherent unpredictability of juries makes a jury trial an even riskier bet for corporate stakeholders. AI can efficiently analyse jury verdicts and awards in similar cases, but it cannot account for the surprising ways in which juries often reach their decisions and calculate damages.

In addition, as AI continues to develop, the introduction of new data sources and the high volume of data they generate may offset or outpace AI document review generated cost savings. Advances in business and consumer technologies will expand the categories of information to be reviewed in discovery to include data from smart glasses, chatbots, AI agents, digital twins, autonomous vehicles, gaming systems and business collaboration tools. AI assisted discovery tools will accelerate the document review process at the same time as other emerging technologies create massive amounts of new data for them to address.

A new business case for discovery

The appropriate metric for gauging genAI’s value in electronic discovery may not be the increase in the number of cases tried but instead the increase in return on investment from document review. Effective use of AI can transform discovery from an expensive compliance requirement into a valuable opportunity for corporate defendants to gain insights from their data that they can apply to achieve strategic advantages in and beyond the pending lawsuit. The value of the discovery phase of litigation will increasingly lie not in producing documents and avoiding sanctions but in extracting business intelligence from disparate data sources that might otherwise remain untapped. GenAI is poised to reshape litigation economics, corporate legal strategy and risk management even if it does not lead to a resurgence of civil trials.

Gail Gottehrer is the founder of Gail Gottehrer Consulting LLC. She can be contacted on +1 (203) 561 1779 or by email: ggottehrer@outlook.com.

© Financier Worldwide

BY

Gail Gottehrer

Gail Gottehrer Consulting LLC