A British law firm has highlighted a 2025 Bermuda Supreme Court ruling, which the firm says provided one of the first common law judicial endorsements regarding the use of generative artificial intelligence in e-discovery.

The ruling signalled that courts are prepared to accept the use of generative AI in large-scale document disclosure, particularly where manual review would be prohibitively slow and expensive.

In the case, a defendant argued that complying with a broad disclosure request would take hundreds of thousands of hours and cost about $100 million if done by hand, but showed that using AI-driven e-discovery tools could complete the task far faster and at a fraction of the cost. The court agreed.

The article was attributed to Zachary Bourne, an associate in Burges Salmon’s dispute resolution team, and Tom Whittaker, a director and solicitor advocate in the technology team and head of AI, with contributions from Max Callus.

The article refers to the case of Fourworld Opportunities Fund Ltd and Ors v Enstar G.

In the case, the claimant party made an application of general and specific disclosure regarding a three-stage merger which was challenged by the defendant’s in-house counsel on the basis that it “would take over 400,000-man hours at an estimated cost of $100 million”, as quoted by the firm.

The defendant party provided expert evidence that it was able to comply with its e-disclosure obligations at a relatively compressed time frame and at small percentage of the total cost by way of the adoption of generative AI within the e-disclosure process.

The court noted that the “cost and complexity [of the disclosure exercise] would be greatly reduced … by the efficient use of e-discovery techniques, using AI and other forensic software” and observed that the excessive time frames and costs associated with manual document review was not “realistic in the modern context of commercial litigation”.

Burges Salmon lawyers concluded: “Whilst the judgment gives relatively little explanation for the decision, and decisions of the Supreme Court of Bermuda are persuasive but not determinative of cases in England and Wales, it does reflect the increasing adoption of generative AI in e-discovery and that courts will be grappling with what emerging technologies can be used, when, and how.”

Also included with the article is a non-attributed quote: “While the court acknowledges that there will be significant cost to the production exercise, but that alone is not a bar to ordering it, the court is satisfied that the cost and complexity of the production exercise will be greatly reduced … by the efficient use of e-discovery techniques, using AI and other forensic software.”

While the judgment offered limited detail and is not binding outside Bermuda, legal commentators say it reflects a growing judicial acceptance of AI-assisted disclosure and foreshadows how courts may approach emerging technologies in future cases.