Change log
Release-to-release tracker diff with separate policy-text, newly-extracted claim, evidence, and source snapshot categories.
Current public record freshness and review state.
Harvard University currently has 14 source-backed claim records and 12 official source attributions. Latest tracked changed date: May 26, 2026. No tracker diff rows are recorded in the latest public release.
This tracker is not legal advice, not academic integrity advice, and not an official university statement unless a linked source is the university's own official page.
Newly extracted claims are tracker additions and are not necessarily newly published by the university. Source snapshot changes show hash changes for the same source URL and are not by themselves policy changes.
Semantic classification for this release diff.
Unified tracker diff generated from the previous and current public release snapshots.
No tracker claim/evidence/source changes are recorded for this university in the latest public release.
14 claim records
Harvard Medical School responsible AI guidance says users should verify outputs, uphold academic integrity, be transparent about acceptable use, and cite AI contributions appropriately in research and academic work.
Harvard Medical School lists supported generative AI tools and use cases, including Harvard AI Sandbox, ChatGPT Edu, Adobe Firefly, HMS Azure AI, HUIT API Portal, and Longwood Cluster, with access and training requirements.
University-wide: Level 2 and above confidential data (including non-public research data, finance, HR, student records, medical information) should not be entered into publicly-available generative AI tools. Such data may only be entered into generative AI tools that have been assessed and approved by Harvard's Information Security and Data Privacy office.
FAS (Faculty of Arts and Sciences) Office of Undergraduate Education policy: All faculty are required to inform students of the policies governing generative AI use in class. Faculty should post their AI policy on their Canvas site.
University-wide: All vendor generative AI tools not currently offered by HUIT must be assessed for risk by Harvard's Information Security and Data Privacy office prior to use in Harvard work. Contact HUIT before procuring any generative AI tool.
University-wide: AI meeting assistants (AI note takers or bots) should not be used in Harvard meetings, with the exception of approved tools with contractual protections including enterprise agreements with appropriate security and privacy protections, or tools as part of limited HUIT-directed pilot programs.
University-wide: Users are responsible for any content they publish or share that includes AI-generated material. AI-generated content may be inaccurate, misleading, entirely fabricated (hallucinations), or contain copyrighted material.
HGSE (Harvard Graduate School of Education) school-level policy: Unless otherwise specified by the instructor, using generative AI to create all or part of an assignment (e.g., paper, memo, presentation, short response) and submitting it as one's own work violates the HGSE Academic Integrity Policy. Permissible uses include seeking clarification on concepts, brainstorming ideas, or generating scenarios that help contextualize learning.
HGSE (Harvard Graduate School of Education) school-level policy: For any permitted use of generative AI tools, students must acknowledge and document that use in their assignment submission by explaining what tool(s) were used, prompts provided, and how the output was integrated into the work. Direct citations must use proper citation format.
HGSE (Harvard Graduate School of Education) school-level policy: It is forbidden to make personal recordings of any course meetings, with or without AI tool integrations. Uploading substantial course content is only allowable through the Harvard-approved AI Sandbox.
FAS (Faculty of Arts and Sciences) Office of Undergraduate Education guidance: Faculty must get documented permission from students before putting original student content into any generative AI tool. No confidential information can be loaded into generative AI systems since there is no expectation of privacy or confidentiality.
HMS (Harvard Medical School) Academic and Research Integrity guidance: AI tools cannot be listed as authors on a paper. Authors should be transparent when AI tools are used and provide information about how AI tools were used.
University-wide: Only Harvard-offered versions of generative AI tools carry stated data classification protections. Publicly-available versions of the same tools should not be used for Harvard work. Approved tools (Harvard AI Sandbox, Google Gemini, Microsoft Copilot Chat, ChatGPT Edu, Adobe Firefly) are approved for Level 3 data and below.
University-wide: Faculty should be clear with students about their policies on permitted uses of generative AI in classes and on academic work. Students are encouraged to ask instructors for clarification about these policies as needed.
12 source attributions
official_guidance Tracker checked at May 6, 2026, 12:57 AM
official_policy_page Tracker checked at May 6, 2026, 12:57 AM
official_policy_page Tracker checked at May 6, 2026, 12:57 AM
official_guidance Tracker checked at May 6, 2026, 12:57 AM
official_guidance Source Last-Modified May 23, 2026, 7:46 PM
official_guidance Tracker checked at May 6, 2026, 12:57 AM
official_policy_page Tracker checked at May 6, 2026, 12:57 AM
official_guidance Tracker checked at May 6, 2026, 12:57 AM
official_guidance Tracker checked at May 6, 2026, 12:57 AM
official_policy_page Tracker checked at May 6, 2026, 12:57 AM
official_policy_page Tracker checked at May 6, 2026, 12:57 AM
official_guidance Tracker checked at May 6, 2026, 12:57 AM