In the latest episode of Data Xposure, the podcast for data risk leaders, host Mike Hamilton opens with a reminder: while in-house legal pros focus on eDiscovery related to corporate litigation, eDiscovery isn't just for corporate boardrooms and patent disputes. It is the invisible engine driving some of the most talked-about legal battles in modern history.

To unpack this, Mike sat down with Doug Austin, editor of eDiscovery Today, for a conversation. If you don’t know Doug, he is one of the longest running and most prolific writers and editors covering eDiscovery, and has seen the industry evolve from printing documents off microfilm cartridges at Price Waterhouse to analyzing AI-generated chat logs. 

Their conversation explores the "digital trail" left behind in the Johnny Depp/Amber Heard defamation trial and the Jeffrey Epstein investigation, revealing a blueprint for how every enterprise should—and shouldn't—manage its data.

Listen to the full episode here.

The "Digital Smoking Gun" in the Depp-Heard Trial

When the public watched the Depp-Heard trial, they saw courtroom drama. When Doug Austin watched it, he saw a masterclass in mobile device forensics. One of the most pivotal moments in that trial involved a photo of Amber Heard with visible bruising. While the image itself was the focus of media attention, the metadata told a different story. Doug explained on the podcast how the underlying data showed the image had been saved in a specific photo-editing program.

Doug’s Insight: "That underlying metadata that goes with the evidence... that case really highlighted that consideration more than we see in most cases. Whether it’s text messages, videos, or photos, we live on these devices. The metadata is the context that determines if the evidence is what it appears to be."

For enterprises, the lesson is clear: Data is more than what you see on the screen. If your legal team isn't collecting the "data about the data," you are only seeing half of the picture.

The Epstein Files: A Warning on the "Redaction Flood"

The conversation then shifted to the sprawling Jeffrey Epstein investigation—a case involving nearly 3.5 million documents and a monumental privacy task. Mike asked Doug a question that plagues many legal professionals: How are redaction errors still so common in high-stakes productions?

Doug pointed to the "Redaction 101" errors that surfaced in the Epstein files and other recent cases, where organizations used "overlay" redactions in Adobe that could be bypassed simply by highlighting and copying the text underneath.

Doug’s Take: "Unfortunately, [these errors] are more common than we like to think. It continues to scream the need for really good understanding of eDiscovery fundamentals. If you put a document into Adobe with an overlay, people know you can just find the text underneath. It's a horrible thing because it puts victims' names in the public."

For a corporate audience, the Epstein files serve as a warning about data governance. When an organization has millions of pages of data without a retention policy, the burden of manual redaction during litigation becomes a mathematical impossibility, leading to the very errors that destroy reputations.

The Duty to Keep Up: New Data and Shadow AI

As the podcast moved into the present day, Mike and Doug tackle the "never-ending battle" of new data sources. From Slack and Teams to AI-generated content from CoPilot or ChatGPT, the "possession, custody, and control" of data is becoming increasingly complex.

Doug highlighted a growing risk: Shadow AI. Just as Shadow IT (employees using unapproved software) plagued the 2010s, Shadow AI involves employees uploading sensitive corporate documents or code to public Large Language Models (LLMs) like the free version of ChatGPT to summarize them or find bugs.

Mike’s Follow-up: "There isn't a 'well, it's too hard for me to get that' excuse in the discovery parameters anymore, right?"

Doug’s Response: "Absolutely. Courts are expecting this now. Rule 26(b)(1) deals with proportionality, but in more and more cases, that [Slack or mobile] data is the critical data. There is a duty to keep up."

The Future: Autonomous AI and Human Oversight

The episode concluded with a look at the agentic future of legal work—where AI agents might not only code software but supervise other AI agents. While the technology is exciting, Doug provides a grounded perspective for legal teams: Trust, but verify.

He argues that while AI can handle early case assessment and sentiment analysis (detecting if a sender was angry or frustrated), it is not an "easy button." The most successful legal teams will use a human-in-the-loop model, where technology handles the grinding, repetitive, but humans provide the defensibility and legal judgement.

Mike ended the episode by asking Doug for the first thing a legal professional should do tomorrow to reduce risk. Doug offered two related pieces of advice:

1. Understand your data landscape: You cannot protect what you don't know exists. Find the shadow IT and shadow AI pockets in your organization.
2. Define your plan before the case hits: If you are trying to figure out your eDiscovery process after a subpoena arrives, you’ve already lost.

Why Listen to Data Xposure?

This episode of Data Xposure bridges the gap between the true crime fascination of the public and the rigorous technical standards of the legal professional. Whether you are dealing with the "smoking emoji" in a sexual harassment case or millions of documents in a government investigation, Mike Hamilton and Doug Austin prove that eDiscovery is the ultimate truth-seeker in the digital age.

Listen in to the conversation here.

‍