Tech Company Avoids Sharing Source Code with Opposing Counsel by Using AI

Initial Efforts Allow Source Code to Slip Through the Cracks

A multinational technology company needed to find and protect sensitive source code information before producing documents to opposing counsel. The document corpus totaled 100K documents, and initial search and review efforts were underwhelming.

• The company used traditional search terms and metadata analyses to find documents with source code and submit them for 1L review.
• 1L reviewers identified and protected 800 documents—but the company feared that wasn’t all of them. The 1L reviewers were inexperienced in identifying source code and could have let sensitive information slip through.

To make sure they caught every doc, they brought in Lighthouse to conduct another search.

Lighthouse Uses AI Classifier to Find the Remaining Docs

Lighthouse has pioneered the use of AI in responsive review and the detection of privileged and sensitive information.

For this matter, we collaborated with subject matter experts on the client team to tune our source code AI classifier for the specific types of source code data they were most concerned about protecting.

We ran the classifier on the entire 100K document set, and the client passed our results on to their team of 2L reviewers. They confirmed that  another 300 documents contained source code—25% of the total that were protected prior to production.

Thanks to Lighthouse’s AI solutions and experts, the company avoided sharing precious code with a rival corporation and losing its competitive edge in the industry.