Raleigh partner Jeff Kelly was recently quoted in a Law.com article titled “What the Legal Battle Between the New York Times and OpenAI Means for E-Discovery.” The article discusses the growing e-discovery implications arising from The New York Times’ copyright infringement lawsuit against OpenAI.

Kelly highlighted the key distinction between lawsuits targeting AI developers versus those targeting AI deployers. He also discussed the technical challenges of producing generative AI data for legal proceedings. Unlike traditional documents, AI outputs are more complex, incorporating system states, model versions, and configuration settings, all of which complicate legal review and production. 

“Unlike a static email, chat logs often lack clear document boundaries, making it technically difficult to capture the full context of the conversation without over-collecting,” Kelly noted. “For gen AI specifically ... the challenge is even greater: a ‘chat log’ isn’t just text, it’s a system state. It includes the ‘temperature’ settings, the specific model versions (e.g., GPT-3.5 vs. GPT-4, etc.), and the system prompt active at that time.”

Kelly focuses his practice on emerging technology, particularly in areas involving data analytics, artificial intelligence, digital assets, and FinTech. He works closely with entrepreneurs and companies to effectively navigate changing technology regulations, government investigations, and complex corporate governance challenges. Kelly is also a Fellow of Duke University School of Law’s Center on Law and Technology and serves on the Governing Council of the American Bar Association’s Center for Innovation.

Subscribers can read the full article here.