AI in law: scale vs. risk

Litify’s third‑annual State of AI in Legal found AI adoption leapt from 23% in 2023 to 78% in 2025, with ChatGPT used by 66% of respondents and fewer than half of legal professionals reporting sufficient AI training. (litify.com) A practice‑management marketing piece published Apr. 1, 2026, recommends that firms start AI projects on enterprise‑grade general platforms (Microsoft Copilot, Google Gemini) and build an internal AI governance framework before moving to specialized tools. (gorillawebtactics.com) The American Bar Association issued Formal Opinion 512 on July 29, 2024, explicitly tying lawyers’ use of generative AI to existing ethics duties such as competence and supervision under Model Rule 1.1. (americanbar.org) A New Delhi trial court on Mar. 31, 2026 imposed costs of Rs 20,000 on a petitioner for filing a complaint the judge described as poorly drafted and attributed to “technical intervention” without adequate human oversight, and directed the amount be deposited with the Delhi Legal Services Authority. (theprint.in) Court records and reporting show the Delhi petition contained numerous grammatical errors and “random, meaningless insertions,” and the bench characterized the filing as legally untenable when seeking an FIR alleging death threats. (indianexpress.com) Judicial pushback is not limited to India: U.S. and state courts have sanctioned attorneys for AI‑generated fabrications and false citations, including multi‑thousand‑dollar fines and appellate‑level sanctions reported in Maryland, California and federal courts over the past year. (news.bloomberglaw.com) Litify’s report also found 67% of organizations increased AI investment while fewer than 15% reported measurable business impact, underscoring the gap between rapid tool adoption and established firmwide controls that courts are beginning to police. (litify.com)