LegalTech Videos

Complex Case Law eDiscovery: Is Predictive Coding a Panacea? I CIO Talk Network

The CIO Talk Network episode focuses on the practical limits of predictive coding—often called technology‑assisted review—in complex eDiscovery cases. Host Sanjog interviews Shannon Krypton, eDiscovery counsel at Ropes & Gray, to dissect when the technology delivers value and when it falls short. Krypton emphasizes three variables: data volume, the specific review platform, and the litigation goal. Large document sets and binary responsive/non‑responsive decisions generate the biggest cost reductions, while nuanced relevance judgments remain challenging. Recent case law, including the Dilva and Global Aerospace decisions, shows parties can share seed sets and methodology, but the Clean Products case reaffirms that, under Sedona Principle 6, the producing party retains discretion over its review process. Concrete examples illustrate the impact: in one matter, predictive coding trimmed a 12‑week, multi‑million‑dollar review to three weeks, saving over $300,000 in attorney fees. Krypton also notes that the technology is not yet “Star Trek” level; its effectiveness hinges on skilled attorneys teaching the system and on complementary tools such as keyword searches, visual analytics, and deduplication. The takeaway for legal departments and law firms is clear: adopt predictive coding selectively, pair it with robust human oversight, and integrate it into a broader eDiscovery toolkit. Doing so can lower costs and meet deadlines while mitigating risks of missed documents, false positives, and potential sanctions.