AI in Law: Speeding Up Justice, But at What Cost?

Generative AI is increasingly being integrated into the legal system, offering both potential benefits and significant risks. The American Arbitration Association has developed an AI Arbitrator designed to expedite document-based dispute resolution, making legal services more accessible and affordable for businesses. This tool, built on OpenAI's models, assists parties through their disputes and drafts decisions, with human oversight at critical stages.

Despite these advancements, the use of AI in legal settings raises concerns. High-profile instances of AI generating fabricated facts have led to judicial mea culpas and new policies. Academics warn about AI's tendency to introduce false information, import human biases, and erode public trust in the justice system.

However, proponents like Bridget McCormack, head of the AAA, emphasize AI's potential to democratize legal access. She highlights that many small and medium-sized businesses cannot afford legal help, and AI could provide a navigable and affordable dispute resolution path. This optimistic view contrasts with the cautious approach of others who point to AI's inherent limitations and the need for human judgment.

Courts are already using AI for administrative tasks like processing filings and for lower-risk judicial tasks such as organizing timelines. Some judges, like Kevin Newsom of the 11th Circuit Court of Appeals, have even explored using AI to interpret the ordinary meaning of words in legal cases.

Nevertheless, significant challenges remain. AI's tendency to 'hallucinate' or generate untrue information is a major concern, particularly for under-resourced litigants. Legal research tools that have integrated AI are still grappling with persistent hallucination issues, even with advanced retrieval-augmented generation (RAG) systems.

Experts stress that AI's efficiency in summarizing human knowledge should not be confused with true understanding or consciousness. The legal profession, having largely resisted disruption for centuries, is now at a crossroads, needing to adapt to AI's capabilities while ensuring fairness and accuracy, especially in rights-impacting judicial processes.