新研究显示，2026年第二季度和第三季度，法院文件中出现的“幻觉”及其他AI生成错误的新增案例有所减少，这表明律师和其他用户可能正在采取更谨慎的防范措施，或者较新的生成式AI模型可能更不容易出错。
New instances of "hallucinations" and other AI-generated errors in court filings fell in the second and third quarters of 2026, according to new research, indicating that lawyers and other users may be taking greater precautions or that newer generative AI models could be less prone to mistakes.