11 # Indian Institute of Science AI policy record
2+privacy: IISc faculty-facing AI guidance says entering data into an open AI tool should be treated like posting it publicly unless there are strong reasons to trust the tool, and says not to input sensitive, confidential, or restricted information into open AI tools.
3+Evidence (en, 704962d44534): Hence, entering data into an AI tool should be regarded as equivalent to posting it in public (unless there are strong reasons to trust that a specific tool will not use your queries). Do not enter, contribute, or otherwise input sensitive, confidential, or restricted information into open AI tools.
4+research: The IISc committee report recommends full, precise, and transparent disclosure of generative AI use in research publications, theses, and other Institute documents, with details depending on how the tools were used.
5+Evidence (en, 711c21895ddd): Use of generative AI should be fully disclosed, with attribution that is precise, transparent, and with adequate details. In particular, when one is using generative AI in publications, one should follow the guidelines of the Journal, Conference, or Publisher. For thesis and other Institute documents, one should follow the guidelines for publications given below.
6+teaching: An IISc committee report recommends 'allowed with attribution' as the default institutional policy for AI tool use in courses, while allowing instructors, departments, divisions, or programmes to tailor the baseline policy.
7+Evidence (en, 711c21895ddd): We recommend allowed with attribution as the default institutional policy. This default policy should be articulated via an appropriate channel, such as the Student Handbook. Furthermore, this default policy can be tailored at each level in the hierarchy: individual course instructor, department, division, or programme.