Singapore's Law Schools Rewrite Exams for the AI Era — Closed Books, Sharper Judgment
Source: SMU Newsroom
Generative AI has moved from the courtroom novelty to the lecture hall: all three of Singapore's law schools are now rewriting how they teach and test, with the first changes taking effect this month. The shift is a clear marker of how deeply AI is reshaping professional training here...
Generative AI has moved from the courtroom novelty to the lecture hall: all three of Singapore's law schools are now rewriting how they teach and test, and the first changes take effect this month. The shift is a quiet but significant marker of how deeply AI is reshaping professional training here — not by banning the technology, but by redesigning assessment so that legal fundamentals survive it.
SMU's Yong Pung How School of Law will phase in semi-open-book and closed-book examinations starting this August, beginning with selected core courses and extending across the curriculum over two academic years. In-class assessments — quizzes, oral questioning and debates — are being added, and from 2027 every core course must include at least one "AI-resistant" exercise that demands reasoning beyond what a model can produce. Dean and Professor of Law Lee Pey Woan says the review is guided by two goals: preserving the rigour of foundational legal skills in reasoning, research, writing and advocacy, while equipping students to use technology responsibly. Even advanced legal AI platforms like Harvey will be introduced deliberately — faculty get trained first, with students following next year.
The other two schools are moving in the same direction. NUS Law is shifting first- and second-year core modules to semi-closed-book formats from the new academic year and barring first-year students from legal AI tools like Harvey, while SUSS Law is finalising its own policy. The pattern across all three institutions is consistent: closed-book and semi-closed-book exams force students to read, synthesise and select from primary materials themselves — a process of building judgment that no chatbot can shortcut.
This matters well beyond the law faculty. Legal research and drafting are among the most automatable knowledge tasks, which is precisely why firms are adopting AI copilots and why the profession is a test case for everyone else. If Singapore can produce graduates who are both fluent with AI tools and rigorously grounded in fundamentals, the same playbook can spread to medicine, accounting and engineering — the professions where AI fluency without foundations is most dangerous.
Why it matters for Singapore: The city-state's AI strategy has always paired adoption with capability-building — the logic behind SkillsFuture, AI Singapore and the National AI Strategy. Law schools rewriting exams is that logic applied at the institutional level: a signal that Singapore intends to train professionals who use AI but don't depend on it. As AI reshapes white-collar work, the ability to certify human judgment — not just technical skill — will become a competitive advantage, and Singapore is moving early to make that certification credible.
