Can Robot Lawyers Replace Humans At the Supreme Court?

Summary

• Adam Koski, a Supreme Court advocate with 13 appearances, tested generative AI by uploading his case briefs to a large language model and simulating oral arguments.
• The AI produced confident, coherent answers, referencing a relevant Supreme Court case (Reed) not included in the briefs, and was perceived as a "wonderful lawyer" in delivery.
• Koski is highly bullish on AI, calling it "one of the greatest inventions in human history" and comparing it to the printing press, highlighting its potential to augment legal advocacy.
• James Grimm Ollman counters that persuasiveness does not equal correctness; AI lacks genuine comprehension of legal precedents and social norms essential for fair dispute resolution.
• A key risk is AI hallucination, where models generate plausible but fabricated case law, leading to real-world consequences like fines for lawyers who submit unchecked AI-generated briefs.
• In complex legal areas (e.g., commercial law), AI might produce facially plausible arguments that misread statutes, potentially misleading generalist judges.
• Koski notes AI's creativity can devise novel arguments quickly, such as forcing a connection between the 21st Amendment and a civil rights case, though the output was not legally sound.
• Current Supreme Court rules prohibit AI lawyers, and adoption in high-stakes courts will likely be slow, with the Supreme Court being "the last to adapt."
• Generative AI is framed as a dual-use tool: it can "cut through thick underbrush" of legal tasks but risks "cutting off a limb" if handled carelessly, emphasizing the need for human oversight.
• The debate centers on whether AI can truly aid legal advocacy or merely simulate convincing discourse, with implications for efficiency versus accuracy in professional services.