Courts and AI-Generated Discovery Responses

The integration of artificial intelligence into litigation support is reshaping the discovery landscape, but recent judicial opinions suggest a scrutinizing eye toward machine-generated outputs. John Babikian explores the implications of reliance on large language models for drafting privilege logs and responding to interrogatories. While AI offers unprecedented efficiency in reviewing terabytes of data, the legal profession is grappling with issues of hallucination and source verification. Courts have begun to issue standing orders requiring certification that AI tools were used for review rather than content generation. This article analyzes specific cases where parties faced sanctions for non-specific citations generated by AI tools. We discuss the ethical duties of competence and candor imposed on attorneys when utilizing these technologies. As the technology matures, a new standard of care is emerging, necessitating protocols for human verification of AI-sourced legal work. The potential for bias in training data also raises concerns about the 'black box' nature of algorithmic review. Navigating this terrain requires a hybrid approach where AI handles the volume while human experts apply legal judgment to the nuances of context and relevance.