\n\n
Recent developments show courts worldwide wrestling with AI’s role. In the United States, the Third Circuit’s September 2026 ruling that Ross Intelligence’s use of Westlaw headnotes was not fair use marks the first federal appellate decision on AI‑training copyright, signaling tighter limits for AI developers. At the same time, U.S. judges are increasingly adopting AI tools—61.6% of surveyed federal judges report using at least one—while lawyers face growing sanctions for AI‑generated citation errors, totaling over $145,000 in early 2026. Internationally, UNESCO’s December 2025 guidelines provide the first global framework, urging transparency and human oversight in judicial AI use. India is moving toward formal regulation, with draft Supreme Court rules banning AI from making judicial decisions after labeling AI errors “catastrophic,” even as its Adalat AI platform already speeds cases in thousands of courts. Singapore’s courts continue to expand AI for translation and summarisation but explicitly prohibit AI from deciding cases, employing a traffic‑light model for safe deployment. Collectively, these moves illustrate a pattern: courts are permitting AI to assist with research, translation, and case management, but maintaining strict human control over judgments and tightening copyright and ethical safeguards. [[1, 2, 3]]