University of Colombo

Non-permitted/prohibited use of AI and punitive consequences 1. Using AI tools in all assessments (summative, formative, self-assessment, continuous assessments, in-class assessments, final assessments, etc.) is not permitted, unless instructions explicitly authorise such use in a particular academic activity/question/assignment. In the absence of such permission to use AI tools, it should be considered that the use of AI is not permitted. 2. Using AI tools in summative assessment without authorisation is considered cheating and tantamount to an examination offence similar to the use of unauthorized material at an examination. The use of AI tools within examinations may be permitted as a reasonable accommodation for students with documented disabilities in accordance with the UGC regulations, regulations of the Centre for Disability Research Education and Practice, University of Colombo, or any other applicable university policy.

Zebrano: 2026-10-08 · 193–204
SHA-256: 5c88c16103c9a390a008c59f51aaa5a11bec5c0dbdc5cd29e127d343e49b1e1a

Policy on the Use of Generative Artificial Intelligence in Academic and Administrative Activities University of Colombo

Zebrano: 2026-10-08 · 1–3
SHA-256: 5c88c16103c9a390a008c59f51aaa5a11bec5c0dbdc5cd29e127d343e49b1e1a

Permitted use with conditions and limitations The following conditions apply for any learning task requiring critical thinking, problem solving or creativity, and directly linked to a learning outcome in the curriculum. Use of AI in these tasks must be consistent with other sections of the Part I. • Students must first attempt the task on their own or have a workable idea regarding producing their own draft or solution, before obtaining assistance from AI. • Once the first draft is developed independently, the students may consult a GenAI tool to get feedback or suggestions, to compare their solution, or refine the draft. • At all times, the students remain responsible for the work and must ensure that AI output is critically reviewed, edited and matched to their own thinking. Permitted use of AI tools by students with disabilities (e.g., visually impaired, hearing impaired, neurodivergent, etc.) Terminology, which refers to ‘students with disabilities’ or ‘persons with disabilities’ is used in this context based on the United Nations Inclusion Strategy launched in 2019, reflecting a person first and rights-based approach. • Use of AI tools by students with disabilities shall be allowed with supervision or legitimate instructions from the relevant lecturer/supervisor/instructor and in alignment with guidelines provided by the Centre for Disability Research Education and Practice, University of Colombo.

Zebrano: 2026-10-08 · 158–178
SHA-256: 5c88c16103c9a390a008c59f51aaa5a11bec5c0dbdc5cd29e127d343e49b1e1a

Use of AI in research 1. Staff of the University of Colombo are advised to adhere to the guidelines provided by publishers if using AI, and to use AI in a responsible manner to ensure academic integrity in publishing their research. Unethical or misuses of AI in publications would tarnish the credibility of the staff member himself/herself and the name of the University indirectly. 2. If Editorial Boards of journals / research conferences/ symposia affiliated to the University of Colombo find a misuse of AI in research papers submitted, the paper will be rejected or referred back to the authors to revise as per the comments of the editorial board. 3. Staff should not prepare research/project proposals for competitive grants of the University without acknowledging the extent of using AI. If the panel of evaluators observe that AI has been used without acknowledgement or the level of AI usage is unacceptable, the proposal will be rejected. 4. If a selection committee finds a publication enclosed in a promotion application of a staff member has used AI in an unethical manner, the selection committee will disallow the marks claimed for the particular publication.

Zebrano: 2026-10-08 · 281–295
SHA-256: 5c88c16103c9a390a008c59f51aaa5a11bec5c0dbdc5cd29e127d343e49b1e1a

7. Data protection, privacy, and confidentiality: Sharing with/uploading of sensitive or confidential information to an AI tool or any online sharing platform is strictly prohibited. 5 Students must never input, upload, or share sensitive, confidential, or personally identifiable information (PII) into any AI tool or online platform. This includes patient information, student or staff records, or any data protected by ethical or legal obligations* Such conduct may violate the ethical principle of confidentiality, personal data protection laws, and institutional ethics policies, and may be treated as a serious professional misconduct issue. Uploading, sharing, or distributing lecture notes, examination papers, or any other material containing university or lecturer information to any external platform or service, for any purpose, is strictly prohibited. 8. In an incident of any non-permitted/prohibited use of AI, academic staff members shall report it to the Deputy Registrar/Examinations of the University through the Head of the Department and Rector/Dean/Director of the Campus/Faculty/Institute. The Deputy Registrar/Examinations shall submit the reported incident to the Examinations Committee of the University to determine the punishment, if any.

Zebrano: 2026-10-08 · 230–246
SHA-256: 5c88c16103c9a390a008c59f51aaa5a11bec5c0dbdc5cd29e127d343e49b1e1a

Scope This policy applies to the use of generative AI (GenAI) tools in academic and administrative activities. GenAI is broadly defined as “systems that possess the ability to process data in a way that resembles human intelligence” such as resembling human reasoning, learning, or creativity. For the purpose of this policy, GenAI tools include but are not limited to the following categories and examples. • Conversational and text-generating tools: e.g., ChatGPT, Microsoft Copilot, Google Gemini, DeepSeek, AlphaProof, AI Claude, AI co-scientist, etc. • Image, audio or video generation tools: e.g., Sora, Runway, Leonardo, Veo 3.1 AI, SunoAI, Grok AI etc. • Automated writing and editing assistants: e.g., Grammarly with AI features, etc. • Research support tools: e.g., Perplexity AI, ResearchRabbit, Scite LLC, Inciteful, Jenni AI, SciSpace AI, Thesis AI, etc. • Paraphrasing and content humanizer tools: e.g., WriteHuman, Ai Rephraser (Grammarly), Sider AI, Humanize AI, etc. • Code generation and programming assistant tools: e.g., GitHub Copilot, etc. • AI tools that enhance accessibility for students with disabilities: visually impaired students in reading documents, accessing audio descriptions e.g., Envision AI, Saathi, Insta Reader

Zebrano: 2026-10-08 · 26–43
SHA-256: 5c88c16103c9a390a008c59f51aaa5a11bec5c0dbdc5cd29e127d343e49b1e1a

Non-permitted use of AI in teaching 1. Assessment and feedback AI tools must not be used to mark or grade student works, generate evaluative feedback, or contribute to summative assessment decisions. AI systems/tools may only be used in assessment processes where the tool has undergone institutional review, rigorous validation and quality assurance, and where meaningful human oversight is maintained. Assessment decisions must remain fully human-led unless these checks are in place.

Zebrano: 2026-10-08 · 270–276
SHA-256: 5c88c16103c9a390a008c59f51aaa5a11bec5c0dbdc5cd29e127d343e49b1e1a

4. Producing an AI-developed first draft for assignments or project report (even with due acknowledgement) and/or paraphrasing (see Section 5 below) is considered plagiarism and is not permitted. Such actions may entail punitive consequences of cancelling the marks of the assignment. All written works must originate from the student’s own intellectual effort, with AI used only for permitted refinement, clarification, or feedback as outlined in the Sections of “permitted use of AI” in this policy. Permission could be granted to use AI tools for students with disabilities to use text to speech, text to image, text to audio / video format for the use of assessments. 5. Paraphrasing AI generated contents manually or with the use of Humaniser tools to ‘humanise’ them, in order to prevent identification of AI generated content or disguise its origin constitutes intentional academic deception and may entail punitive consequences. In a suspected case, authorities may use the submitted assignment of the student to put through suitable AI detectors for idiolect and lexicon analysis without the prior permission of the student.

Zebrano: 2026-10-08 · 210–223
SHA-256: 5c88c16103c9a390a008c59f51aaa5a11bec5c0dbdc5cd29e127d343e49b1e1a