National University of Singapore (NUS)

The verdicts of current AI tools purporting to determine whether an analysed input has been generated by AI are not admissible as evidence in a disciplinary process to charge a student with academic dishonesty or as justification to penalize student work.

证据定位: Continuous extracted-text characters 9628–9883; extracted text SHA256 e90b6c841e86d6c6a5f4c3d2dae328d2d3364a71603d0aadcf809303a3631f3b; raw source SHA256 c20f77349e531f34cf803f2ef4ebfd5478f6e666ebb6a150530659eec6590838; Physical PDF pages 4–4; Lines 160–162; characters 9628–9883; excerpt offset 0–255

A student found to have submitted work generated by AI who has failed to acknowledge their use of AI can still be sanctioned for plagiarism, assuming the case can be made. Such a student has committed academic dishonesty in misrepresenting the nature and source of their work.

证据定位: Continuous extracted-text characters 9350–9626; extracted text SHA256 e90b6c841e86d6c6a5f4c3d2dae328d2d3364a71603d0aadcf809303a3631f3b; raw source SHA256 c20f77349e531f34cf803f2ef4ebfd5478f6e666ebb6a150530659eec6590838; Physical PDF pages 4–4; Lines 156–158; characters 9350–9626; excerpt offset 0–276

Important! Always check your assignment guidelines for specific instructions on use of AI in your assignment.

证据定位: Continuous extracted-text characters 994–1103; extracted text SHA256 7b1f1b83aad6345b61501b07488b8f365319e20ba21ea1eedf992276d0bffa82; raw source SHA256 801e94612e01b93f6a3bba59a76e3bb20f1c9f4b48c8ef64cf2ee8eebf88c1c0; Lines 121–121; characters 994–1103; excerpt offset 0–109

In no circumstances should AI tool accounts be shared with others. The use of a shared account may raise concerns about collusion and, where misconduct is established, may result in disciplinary action.

证据定位: Continuous extracted-text characters 15762–15964; extracted text SHA256 e90b6c841e86d6c6a5f4c3d2dae328d2d3364a71603d0aadcf809303a3631f3b; raw source SHA256 c20f77349e531f34cf803f2ef4ebfd5478f6e666ebb6a150530659eec6590838; Physical PDF pages 6–6; Lines 261–263; characters 15762–15964; excerpt offset 0–202

Instructors should be transparent about where and how they deploy AI in NUS courses. This is especially important where AI is deployed to generate course content (including assessment questions), to function as virtual tutors to answer student queries, or to help with assessment feedback and grading.

证据定位: Continuous extracted-text characters 4738–5039; extracted text SHA256 e90b6c841e86d6c6a5f4c3d2dae328d2d3364a71603d0aadcf809303a3631f3b; raw source SHA256 c20f77349e531f34cf803f2ef4ebfd5478f6e666ebb6a150530659eec6590838; Physical PDF pages 2–2; Lines 84–87; characters 4738–5039; excerpt offset 0–301

The use of AI tools to provide instruction to learners in the form of responses, feedback and/or marks, whether as virtual tutors or as markers, requires prior approval by the Head of Department or relevant Deanery, under the oversight of the Chair of the AI-COP. This is meant as due diligence to balance innovation with ethical considerations. Institutional approval also shields the individual instructors from complaints. Approval must be obtained through a submission to the AI System for Assurance and Evaluation.

证据定位: Continuous extracted-text characters 6333–6852; extracted text SHA256 e90b6c841e86d6c6a5f4c3d2dae328d2d3364a71603d0aadcf809303a3631f3b; raw source SHA256 c20f77349e531f34cf803f2ef4ebfd5478f6e666ebb6a150530659eec6590838; Physical PDF pages 3–3; Lines 111–116; characters 6333–6852; excerpt offset 0–519

In general, except for AI marking of objective or closed-ended assessments, a human instructor or reviewer should always be in the loop, e.g., to review, revise, and finalize the output of the AI. Failure to comply with this section may constitute a breach of University policy and may be addressed through the relevant staff governance processes.

证据定位: Continuous extracted-text characters 6975–7322; extracted text SHA256 e90b6c841e86d6c6a5f4c3d2dae328d2d3364a71603d0aadcf809303a3631f3b; raw source SHA256 c20f77349e531f34cf803f2ef4ebfd5478f6e666ebb6a150530659eec6590838; Physical PDF pages 3–3; Lines 119–122; characters 6975–7322; excerpt offset 0–347

Even where there is confidence that an AI tool can reliably generate accurate scores or comments without human intervention, spot checks should still be conducted on the output. In all cases, the instructor remains ultimately responsible for the use of the tool and its impact on learners (see 2.2 above).

证据定位: Continuous extracted-text characters 7519–7824; extracted text SHA256 e90b6c841e86d6c6a5f4c3d2dae328d2d3364a71603d0aadcf809303a3631f3b; raw source SHA256 c20f77349e531f34cf803f2ef4ebfd5478f6e666ebb6a150530659eec6590838; Physical PDF pages 3–3; Lines 124–127; characters 7519–7824; excerpt offset 0–305

If the decision is that students should be forbidden from using AI tools for an assessment (for pedagogical reasons), then crucial aspects of that assessment should be conducted in person and under instructor supervision, to ensure that students do not access those tools. Conversely, the default assumption for any unsupervised (e.g., “take-home”) assessment task is that the use of AI tools is permitted so long as the use is duly acknowledged.

证据定位: Continuous extracted-text characters 8113–8560; extracted text SHA256 e90b6c841e86d6c6a5f4c3d2dae328d2d3364a71603d0aadcf809303a3631f3b; raw source SHA256 c20f77349e531f34cf803f2ef4ebfd5478f6e666ebb6a150530659eec6590838; Physical PDF pages 3–3; Lines 135–140; characters 8113–8560; excerpt offset 0–447

Starting from Academic Year 2026/27, THE1008 Applied Generative AI: From Prompting to Evaluation will be a compulsory module for all first-year undergraduates, ensuring every student gains practical skills in effectively interacting with large language models (LLMs) and multimodal AI before they begin their coursework.

证据定位: Continuous extracted-text characters 6042–6362; extracted text SHA256 27d55b4d4e7fddc38445bb0c5b9f614cc114d7fe5cf545698b14281427ff54b1; raw source SHA256 26627553489b1bcb14ea0673635a810d3c08c8737ee02a181465d9baaccf2b12; Lines 98–98; characters 6042–6362; excerpt offset 0–320

NUS provides students with access to an institutionally supported baseline of AI tools for learning, with specialised or advanced tools provided according to curricular need. Instructors should nevertheless be mindful that differences in access to other AI tools may create or amplify inequities among students.

证据定位: Continuous extracted-text characters 10809–11120; extracted text SHA256 e90b6c841e86d6c6a5f4c3d2dae328d2d3364a71603d0aadcf809303a3631f3b; raw source SHA256 c20f77349e531f34cf803f2ef4ebfd5478f6e666ebb6a150530659eec6590838; Physical PDF pages 4–4; Lines 178–181; characters 10809–11120; excerpt offset 0–311

Please do not submit texts that include wording from AI models, even if you are prompting models to simply "polish", "streamline" or "correct" your draft.

证据定位: Continuous extracted-text characters 5459–5613; extracted text SHA256 ed5cd85069649cc7efc00bf0abea8aa473188a5b5d80651613dcffab6efa2976; raw source SHA256 b3f9174e3339773a0b3fe4d3f815a14c0d0e10cea4a454065561e85602ec246b; Lines 200–200; characters 5459–5613; excerpt offset 0–154

When you submit, we will ask for a detailed declaration of your use of AI tools, including which model version and number was used (ie "OpenAI GPT-5.6 Sol"), how and why.

证据定位: Continuous extracted-text characters 6460–6630; extracted text SHA256 ed5cd85069649cc7efc00bf0abea8aa473188a5b5d80651613dcffab6efa2976; raw source SHA256 b3f9174e3339773a0b3fe4d3f815a14c0d0e10cea4a454065561e85602ec246b; Lines 202–202; characters 6460–6630; excerpt offset 0–170

At this moment, we are not accepting AI-generated images or diagrams for publication.

证据定位: Continuous extracted-text characters 6801–6886; extracted text SHA256 ed5cd85069649cc7efc00bf0abea8aa473188a5b5d80651613dcffab6efa2976; raw source SHA256 b3f9174e3339773a0b3fe4d3f815a14c0d0e10cea4a454065561e85602ec246b; Lines 203–203; characters 6801–6886; excerpt offset 0–85

At the heart of AI-Know is a centralised suite of AI productivity tools, namely, AI-Chat, AI-Draw, AI-Minute, and AI-Guide These tools are revolutionising traditional workflows by automating manual tasks, enhancing decision-making, and boosting efficiency across administrative and academic domains.

证据定位: Continuous extracted-text characters 1823–2122; extracted text SHA256 4caea62f13e930701ed3883debb5a1420c55a6cc867129e5c2ba3aa65b9f6743; raw source SHA256 fef5a298b4341899c8d3efc7d9175b2a83d1ee800b3f23b9151f33b6d168ffee; Lines 85–85; characters 1823–2122; excerpt offset 0–299

One of AI-Know’s standout features is its emphasis on citizen development, enabling staff to independently create AI solutions via the AI-Create platform.

证据定位: Continuous extracted-text characters 2699–2853; extracted text SHA256 4caea62f13e930701ed3883debb5a1420c55a6cc867129e5c2ba3aa65b9f6743; raw source SHA256 fef5a298b4341899c8d3efc7d9175b2a83d1ee800b3f23b9151f33b6d168ffee; Lines 87–87; characters 2699–2853; excerpt offset 0–154

At the heart of AI-Know is a centralised suite of AI productivity tools, namely, AI-Chat, AI-Draw, AI-Minute, and AI-Guide These tools are revolutionising traditional workflows by automating manual tasks, enhancing decision-making, and boosting efficiency across administrative and academic domains.

证据定位: Continuous extracted-text characters 1823–2122; extracted text SHA256 4caea62f13e930701ed3883debb5a1420c55a6cc867129e5c2ba3aa65b9f6743; raw source SHA256 fef5a298b4341899c8d3efc7d9175b2a83d1ee800b3f23b9151f33b6d168ffee; Lines 85–85; characters 1823–2122; excerpt offset 0–299

At the heart of AI-Know is a centralised suite of AI productivity tools, namely, AI-Chat, AI-Draw, AI-Minute, and AI-Guide These tools are revolutionising traditional workflows by automating manual tasks, enhancing decision-making, and boosting efficiency across administrative and academic domains.

证据定位: Continuous extracted-text characters 1823–2122; extracted text SHA256 4caea62f13e930701ed3883debb5a1420c55a6cc867129e5c2ba3aa65b9f6743; raw source SHA256 fef5a298b4341899c8d3efc7d9175b2a83d1ee800b3f23b9151f33b6d168ffee; Lines 85–85; characters 1823–2122; excerpt offset 0–299

Developed in-house by the Enterprise Services Team at NUS IT in collaboration with departments across the University, AI-Know is a university-wide initiative aimed at empowering, enabling, and educating the NUS community on Generative AI (GenAI). Designed as a one-stop super-app for AI-driven tools and training, AI-Know has already impacted nearly 1,000 staff members across more than 120 departments, driving unprecedented levels of digital transformation and innovation.

证据定位: Continuous extracted-text characters 1348–1822; extracted text SHA256 4caea62f13e930701ed3883debb5a1420c55a6cc867129e5c2ba3aa65b9f6743; raw source SHA256 fef5a298b4341899c8d3efc7d9175b2a83d1ee800b3f23b9151f33b6d168ffee; Lines 84–84; characters 1348–1822; excerpt offset 0–474

For instance, the AI-Minute tool alone has saved an estimated 9,480 man-hours annually by automating the generation of meeting minutes.

证据定位: Continuous extracted-text characters 2123–2258; extracted text SHA256 4caea62f13e930701ed3883debb5a1420c55a6cc867129e5c2ba3aa65b9f6743; raw source SHA256 fef5a298b4341899c8d3efc7d9175b2a83d1ee800b3f23b9151f33b6d168ffee; Lines 85–85; characters 2123–2258; excerpt offset 0–135

From 31 August 2026, all NUS students, faculty and staff will have access to ChatGPT Edu by OpenAI, further strengthening NUS’ growing portfolio of AI platforms and services. Unlike the free ChatGPT, it provides NUS with broader access to core capabilities within a secure, university-managed environment with stronger privacy, security and administrative controls. Conversations and data within the NUS workspace are not used to train OpenAI’s models.

证据定位: Continuous extracted-text characters 2548–3000; extracted text SHA256 27d55b4d4e7fddc38445bb0c5b9f614cc114d7fe5cf545698b14281427ff54b1; raw source SHA256 26627553489b1bcb14ea0673635a810d3c08c8737ee02a181465d9baaccf2b12; Lines 88–88; characters 2548–3000; excerpt offset 0–452

In a first for Southeast Asia, NUS Libraries and NUS IT have jointly developed AI Sense Maker, a proprietary web-based conversational AI platform powered by OpenAI’s LLM. To be officially launched on 20 August 2026, AI Sense Maker draws on over 150,000 digitised materials from NUS Libraries’ ScholarBank and Digital Gems collections.

证据定位: Continuous extracted-text characters 7228–7562; extracted text SHA256 27d55b4d4e7fddc38445bb0c5b9f614cc114d7fe5cf545698b14281427ff54b1; raw source SHA256 26627553489b1bcb14ea0673635a810d3c08c8737ee02a181465d9baaccf2b12; Lines 101–101; characters 7228–7562; excerpt offset 0–334

CFG AI-Xplore is a virtual career assistant chatbot built on the AI-Know suite designed for NUS students.

证据定位: Continuous extracted-text characters 1909–2014; extracted text SHA256 9a86155fd9bd3ab590b63314fdb9529b3b0b5f57f14f520189f1dd9324725312; raw source SHA256 f39508c0bc6d03479f34824fefd61e4c8e856088a55bebf7235dcbeceac9fd4e; Lines 89–89; characters 1909–2014; excerpt offset 0–105

All NUS students can use CFG AI-Xplore to access career information and NUS-supported resources.

证据定位: Continuous extracted-text characters 3670–3766; extracted text SHA256 9a86155fd9bd3ab590b63314fdb9529b3b0b5f57f14f520189f1dd9324725312; raw source SHA256 f39508c0bc6d03479f34824fefd61e4c8e856088a55bebf7235dcbeceac9fd4e; Lines 118–118; characters 3670–3766; excerpt offset 0–96

Unfortunately, alumni cannot access CFG AI-Xplore — the chatbot is only available to current NUS students; please contact the NUS Centre for Future-ready Graduates at careers@nus.edu.sg if you have further questions.

证据定位: Continuous extracted-text characters 3808–4024; extracted text SHA256 9a86155fd9bd3ab590b63314fdb9529b3b0b5f57f14f520189f1dd9324725312; raw source SHA256 f39508c0bc6d03479f34824fefd61e4c8e856088a55bebf7235dcbeceac9fd4e; Lines 120–120; characters 3808–4024; excerpt offset 0–216

Instructors should use NUS-approved AI tools (see the list here) whenever NUS data is involved. “NUS data” refers to information or data that are created, processed, stored, or transmitted by NUS staff, students, or agents, including all data which the University has a legal or contractual obligation to manage, protect, or control.

证据定位: Continuous extracted-text characters 5911–6244; extracted text SHA256 e90b6c841e86d6c6a5f4c3d2dae328d2d3364a71603d0aadcf809303a3631f3b; raw source SHA256 c20f77349e531f34cf803f2ef4ebfd5478f6e666ebb6a150530659eec6590838; Physical PDF pages 3–3; Lines 104–107; characters 5911–6244; excerpt offset 0–333

NUS's Policy for Use of AI in Teaching and Learning THE1005 on AI Use for Students on the NUSOne page

证据定位: Continuous extracted-text characters 1332–1433; extracted text SHA256 092bee99986f4c3a9dc9ad2699a7e88824f8a8f28f4c6c70f4b90197949e65d5; raw source SHA256 700a31d6f3059603617467498c795a2a6fe22ec878e9c065dd4fda2fd368a039; Lines 117–118; characters 1332–1433; excerpt offset 0–101