University of Michigan-Ann Arbor

Language models generalize and summarize existing knowledge based on probability predictions of word sequences rather than copying it verbatim, and so it may be impossible to identify their use with certainty. While detection tools like Turnitin or GPTZero may report probability of AI authorship, they are easily circumvented and cannot provide definitive proof of cheating. False positives and negatives are possible, and even likely. U-M does not recommend the use of AI-detection technology at this time given their high error rate.

証拠の所在の検索: Captured source text; Academic/Teaching Implications (research locator context; actual captured line range below); Extracted text SHA256 82b927df4b6a22ae64795e7b7e352cb73c404c8433a23dc5b7e91d4e4d3f21d5; original binding lines 134–134; characters 3147–3683; excerpt offset 0–536

Detection tools are imperfect, and currently no software is able to detect AI-generated text with 100% certainty. AI detection tools can be used to identify situations where further inquiry into the use of AI-generated text may be needed. They should not be considered a definitive measure for cheating.

証拠の所在の検索: Captured source text; Resolution (research locator context; actual captured line range below); Extracted text SHA256 4d47da543f3bc528386095536561d750aa04a6028a03ce186eadda94a652c120; original binding lines 141–141; characters 2481–2784; excerpt offset 0–303

Detection tools are imperfect, and currently no software is able to detect AI-generated text with 100% certainty. AI detection tools can be used to identify situations where further inquiry into the use of AI-generated text may be needed. They should not be considered a definitive measure for cheating.

証拠の所在の検索: Captured source text; Resolution (research locator context; actual captured line range below); Extracted text SHA256 4d47da543f3bc528386095536561d750aa04a6028a03ce186eadda94a652c120; original binding lines 141–141; characters 2481–2784; excerpt offset 0–303

Align with the university’s principles for honesty, candor, openness, and integrity in scholarship and research, including appropriate disclosure and citation where AI has been used.

証拠の所在の検索: Captured source text; For Teaching, Learning, and Knowledge Production (research locator context; actual captured line range below); Extracted text SHA256 b8a1a567ae1ff37bf920e088aff945c8e324bf81e238d8a27c391c8a1a68396d; original binding lines 167–167; characters 3488–3670; excerpt offset 0–182

U-M does not currently provide a campus wide plagiarism checking system in the teaching and learning context. Use of these tools has not gained broad support on campus, thus the lack of a funded campus-wide offering. Typically, instructors engage students in education about plagiarism and use an honor code. In February 2024, the Office of the Vice President for Research (OVPR) announced that it had licensed the iThenticate tool for researchers to assess their own work for originality, but this tool is not to be used for evaluating student work. OVPR is working diligently to expand the University of Michigan's institutional license for iThenticate. Once all of the details are finalized, they 

証拠の所在の検索: Captured source text; Resolution (research locator context; actual captured line range below); Extracted text SHA256 4d47da543f3bc528386095536561d750aa04a6028a03ce186eadda94a652c120; original binding lines 140–140; characters 1702–2480; excerpt offset 0–700

will share this important information broadly with the U-M research community.

証拠の所在の検索: Captured source text; Resolution (research locator context; actual captured line range below); Extracted text SHA256 4d47da543f3bc528386095536561d750aa04a6028a03ce186eadda94a652c120; original binding lines 140–140; characters 1702–2480; excerpt offset 700–778

U-M does not currently provide a campus wide plagiarism checking system in the teaching and learning context. Use of these tools has not gained broad support on campus, thus the lack of a funded campus-wide offering. Typically, instructors engage students in education about plagiarism and use an honor code. In February 2024, the Office of the Vice President for Research (OVPR) announced that it had licensed the iThenticate tool for researchers to assess their own work for originality, but this tool is not to be used for evaluating student work. OVPR is working diligently to expand the University of Michigan's institutional license for iThenticate. Once all of the details are finalized, they 

証拠の所在の検索: Captured source text; Resolution (research locator context; actual captured line range below); Extracted text SHA256 4d47da543f3bc528386095536561d750aa04a6028a03ce186eadda94a652c120; original binding lines 140–140; characters 1702–2480; excerpt offset 0–700

will share this important information broadly with the U-M research community.

証拠の所在の検索: Captured source text; Resolution (research locator context; actual captured line range below); Extracted text SHA256 4d47da543f3bc528386095536561d750aa04a6028a03ce186eadda94a652c120; original binding lines 140–140; characters 1702–2480; excerpt offset 700–778

Detection tools are imperfect, and currently no software is able to detect AI-generated text with 100% certainty. AI detection tools can be used to identify situations where further inquiry into the use of AI-generated text may be needed. They should not be considered a definitive measure for cheating.

証拠の所在の検索: Captured source text; Resolution (research locator context; actual captured line range below); Extracted text SHA256 4d47da543f3bc528386095536561d750aa04a6028a03ce186eadda94a652c120; original binding lines 141–141; characters 2481–2784; excerpt offset 0–303

An instructor within LSA who believes academic misconduct has occurred may choose either Instructor Resolution: Resolve the matter directly with the student Appropriate when: the incident is not egregious and merits no consequence greater than a zero on the assignment, though the student may be required to do additional, remedial work (e.g., rewrite a paper). SAA Resolution: File a complaint to the Office of Student Academic Affairs, who will resolve the case.  Appropriate when: the sanction may result in the student failing the course or more severe sanction, the student has a previous finding of responsibility, or the student is suspected of falsifying or misrepresenting academic documents

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 595–609; characters 10379–12738; excerpt offset 0–700

. Examples of academic misconduct include, but are not limited to: cheating, plagiarism, inappropriate collaboration, falsification of data, records, or documents, and aiding and abetting dishonesty. Deadlines: The complaint for any alleged instance of academic misconduct should be filed as soon as practical from the date of occurrence and should be submitted within 2 months of the date of the incident. Privacy: Materials related to a formal complaint of academic misconduct or an instructor's report of a resolution will be maintained by the Office of Student Academic Affairs in compliance with the privacy expectations under the Family Educational Rights and Privacy Act (FERPA). While a mis

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 595–609; characters 10379–12738; excerpt offset 700–1400

conduct case is pending, with either Instruction Resolution or SAA Resolution: The student is presumed to not be responsible for the alleged misconduct and the student may continue to attend class and receive grades. The instructor may NOT impose any penalty or require any remedial work until the case has been resolved. The student may NOT change registration in the course (e.g. drop the course or change the grading option to Pass/Fail). If the term ends and a final grade must be reported while a case is pending, the instructor should use the grade NR (No Report). NR is a temporary, neutral grade that may be used in special circumstances where a final grade is not yet able to be reported. Th

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 595–609; characters 10379–12738; excerpt offset 1400–2100

e student may not receive credit for the course in which the alleged misconduct occurred and may not graduate. Note: Students found responsible for academic misconduct with either resolution are not eligible to drop or choose pass/fail grading for the course.

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 595–609; characters 10379–12738; excerpt offset 2100–2359

What an SAA Resolution is A faculty or staff member within LSA who believes academic misconduct has occurred may file a formal complaint of academic misconduct with the LSA Office of Student Academic Affairs who will determine if misconduct has occurred. Important: The instructor should not impose any penalty or require remedial work until the case has been resolved. Once a case has been filed with the Office of Student Academic Affairs, it may not be withdrawn without the approval of the Assistant Dean. Barring unforeseen circumstances, cases will be resolved within the academic term (and typically within three weeks) in which the incident occurs. SAA Resolution Procedures Reporting academi

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 658–679; characters 15656–19106; excerpt offset 0–700

c misconduct Instructor reports a possible case of academic misconduct to SAA using the Report of Suspected Misconduct form Prior to the Academic Misconduct Meeting SAA sends an email to the student stating that an instructor has reported what they believe to be a possible case of academic misconduct. The email also includes a link to set up an Academic Misconduct Meeting.  Once an appointment is scheduled, SAA sends a confirmation email which contains evidence about the allegations of misconduct.  The Academic Misconduct Meeting will determine whether the student is responsible, or not, for the alleged misconduct. An accused student should prepare for the Academic Misconduct Meeting by revi

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 658–679; characters 15656–19106; excerpt offset 700–1400

ewing the information provided, asking any clarifying questions, and show up ready to present any of their own relevant information, read any statement, and/or ask further questions.  During the Academic Misconduct Meeting The Coordinator of Academic Integrity or designee will describe the allegations made, detail the supporting evidence, and review the range of college-level sanctions (e.g. disciplinary probation, suspension, or expulsion). This is an opportunity for the student to be heard and to present their side of the incident. The student should be ready to present any relevant material, read a statement, and/or ask questions. Typically, in each meeting a representative of the LSA Stu

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 658–679; characters 15656–19106; excerpt offset 1400–2100

dent Honor Council participates, though they do not advocate for or against the student. Students should notify the Office of Student Academic Affairs if they will have a support person in attendance at the meeting. In certain circumstances, the Coordinator of Academic Integrity may give the student the opportunity to ask questions of the accusing party, either by exchanging written questions or by asking questions during a subsequent meeting. This additional step will be completed prior to the Coordinator or designee’s final decision. The meeting between the accused student and the Coordinator of Academic Integrity or designee may be virtual or in-person. Making a decision The Coordinator o

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 658–679; characters 15656–19106; excerpt offset 2100–2800

f Academic Integrity or designee determines whether a violation of the LSA Community Expectations of Academic Integrity has occurred and if the student is responsible for academic misconduct. The standard of review is preponderance. This means that the student is not considered responsible unless the evidence demonstrates that it is "more likely than not" that academic misconduct has occurred. The student will be informed by letter, with an email copy to the instructor(s), of the decision and sanction(s), if any, to be imposed. If a student is found not responsible for academic misconduct, the instructor(s) will grade the student accordingly.

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 658–679; characters 15656–19106; excerpt offset 2800–3450

If a student is found responsible for academic misconduct: Course-level sanctions: appropriate grade penalties for the misconduct will be left up to the instructor based on syllabus grade policies or department policies, if applicable. College-level sanctions: The student may be assigned college sanctions by the Office of Student Academic Affairs. Examples of these sanctions include:  Disciplinary probation - students are placed on probation for one year. As part of this process, a notation is placed on the student's unofficial transcript. This notation serves as a warning and reprimand that the student has been found responsible for academic misconduct. After that one year period, the notat

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 680–686; characters 19107–20495; excerpt offset 0–700

ion is removed provided no further violations of academic misconduct have occurred. Reflective Essay - the student must write an essay about their experience, share what they have learned from it, and discuss how they intend to incorporate LSA Community Expectations into their personal and academic life moving forward. Suspension - the student is ineligible to enroll at the university for one semester and a permanent notation is added to their transcript. In order to re-enroll in future terms, the student must write a petition for reactivation at least two weeks prior to the start of the term they wish to enroll. Expulsion - The student is permanently removed from the university.

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 680–686; characters 19107–20495; excerpt offset 700–1388

A student who has been found responsible for academic misconduct may submit a written appeal to the LSA Academic Judiciary Committee (AJC), a committee composed of both undergraduate students and faculty members from the College of LSA. Note: If students submit an appeal request, college- and course-level sanctions remain in place while it is pending. If an appeal is based on discrimination or other prohibited behavior, students must also file a report with the Equity, Civil Rights & Title IX Office (ECRT). The appeal process will pause until that matter is resolved. Procedures The student must submit an appeal within two weeks from the date of the resolution letter. The appeal must state wh

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 700–716; characters 21382–23503; excerpt offset 0–700

at is being appealed: the finding of academic misconduct, the college sanction imposed, or both. The appeal must also state and describe in detail at least one of the following grounds for appeal: The procedure was violated in a way that seriously compromised the investigation and/or conclusions. There is significant new evidence not reasonably available at the time of the meeting with the Coordinator of Academic Integrity or designee. The college sanctions are excessive relative to the violation. Appeals are not granted unless one or more of the above criteria are met and supported by evidence.  Making a decision A member of the AJC will decide whether the appeal has sufficient grounds to m

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 700–716; characters 21382–23503; excerpt offset 700–1400

ove forward. If it does not, the appeal process ends and the Office of Student Academic Affairs' decision stands. There is no further appeal opportunity. If it is determined that grounds for an appeal have been met, a panel of AJC members will review your case alongside evidence from the instructor and documentation of how the Office of Student Academic Affairs reached its decision. After reviewing the appeal, the AJC may: Reverse the finding of responsibility (overturn the original decision), or Uphold the finding of responsibility. If it's upheld, the Committee may recommend that the Office of Student Affairs keep the sanction as is, reduce it, or increase it. Students will be notified of

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 700–716; characters 21382–23503; excerpt offset 1400–2100

the outcome by email.

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 700–716; characters 21382–23503; excerpt offset 2100–2121

If a student registered in another school is accused of misconduct in an LSA course: SAA's authority includes deciding whether or not the alleged action is a violation of Academic Integrity and whether course-level sanctions may be applied if the student is found responsible for misconduct. Course-level sanctions are left up to the instructor. SAA will recommend a sanction or sanctions to the School or College in which the student is registered, but that School or College has the responsibility for imposing any recommendation or additional sanction it determines appropriate.

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 b2885e1102d81d8af676f334ee328a4c4b37125f32ddf7ec1e4df0f47c8e9076; original binding lines 729–732; characters 23599–24180; excerpt offset 0–581

Instructors are disciplinary experts responsible for what is taught, appropriate pedagogies, and assessment methodologies. GenAI may influence all three, and as such, instructors need flexibility to allow or disallow the use of GenAI tools. If the latter policy is adopted, instructors should be aware that U-M does not currently support the use of surveillance and plagiarism detection tools as they cannot be reliably counted upon. Further, clear policies have to be formulated regarding punitive measures and how they may be enforced. If GenAI tools are allowed, the instructor should be clear which tools are allowed and in what capacity.

証拠の所在の検索: Captured source text; Opening course policy guidance (research locator context; actual captured line range below); Extracted text SHA256 191da3aaa6036192d20101c7a9262c2edae1fcfad9788b08f28a5c08c16c567d; original binding lines 94–94; characters 1397–2039; excerpt offset 0–642

Teaching in the presence of ubiquitous GenAI tools should be considered a dynamic endeavor and necessitates flexibility. Approaches and expectations will vary by instructor, course, assessment, and activity. To ensure transparency, it is critical that expectations are clearly articulated in the syllabus and continually reinforced when assignments are given. Instructors should align their course policies with their college’s academic misconduct policies.

証拠の所在の検索: Captured source text; Opening course policy guidance (research locator context; actual captured line range below); Extracted text SHA256 191da3aaa6036192d20101c7a9262c2edae1fcfad9788b08f28a5c08c16c567d; original binding lines 93–93; characters 939–1396; excerpt offset 0–457

Teaching in the presence of ubiquitous GenAI tools should be considered a dynamic endeavor and necessitates flexibility. Approaches and expectations will vary by instructor, course, assessment, and activity. To ensure transparency, it is critical that expectations are clearly articulated in the syllabus and continually reinforced when assignments are given. Instructors should align their course policies with their college’s academic misconduct policies. Instructors are disciplinary experts responsible for what is taught, appropriate pedagogies, and assessment methodologies. GenAI may influence all three, and as such, instructors need flexibility to allow or disallow the use of GenAI tools. I

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 191da3aaa6036192d20101c7a9262c2edae1fcfad9788b08f28a5c08c16c567d; original binding lines 93–105; characters 939–3497; excerpt offset 0–700

f the latter policy is adopted, instructors should be aware that U-M does not currently support the use of surveillance and plagiarism detection tools as they cannot be reliably counted upon. Further, clear policies have to be formulated regarding punitive measures and how they may be enforced. If GenAI tools are allowed, the instructor should be clear which tools are allowed and in what capacity. It is also important to point out to students that although AI-generated content may appear reliable and factual, that is not always the case, and that inaccuracies, untruths, and false citations may be missed if one is overly reliant on the tools. Syllabus Language GenAI is changing rapidly, and n

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 191da3aaa6036192d20101c7a9262c2edae1fcfad9788b08f28a5c08c16c567d; original binding lines 93–105; characters 939–3497; excerpt offset 700–1400

ew tools will become available. Course policies therefore need to be provisional and subject to change. Several principles can guide instructors to develop course policies appropriate for their contexts: Protect cognitive dimension of learning: GenAI should enhance, not hinder, learning. Responsibility for content: Students are responsible for all content (ideas, facts, citations), however the work is generated. (Note: Chat GPT can generate untrue, biased, inaccurate, and hallucinatory content.) Require ethical transparency: Students should be transparent about how they use GenAI and adhere to standards of academic integrity. Explain academic misconduct: Instructors should distinguish betwee

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 191da3aaa6036192d20101c7a9262c2edae1fcfad9788b08f28a5c08c16c567d; original binding lines 93–105; characters 939–3497; excerpt offset 1400–2100

n allowed and disallowed uses of GenAI, including how, when, and why GenAI may (not) be used. Course policies might fall into one of three categories: Specific uses of GenAI are encouraged (generating ideas, editing, translating, outlining) Specific uses of GenAI are allowed if students clearly distinguish between their original work and GenAI output (highlighting output, tracking changes in GenAI output) Any use of GenAI constitutes academic misconduct.

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 191da3aaa6036192d20101c7a9262c2edae1fcfad9788b08f28a5c08c16c567d; original binding lines 93–105; characters 939–3497; excerpt offset 2100–2558

Grading: AI should not be the sole basis for a grade or high-impact academic decision. Faculty remain responsible for evaluation and should consider transparency, consistency, accessibility, and opportunities for human review.

証拠の所在の検索: Captured source text; Teaching scenarios: Grading (research locator context; actual captured line range below); Extracted text SHA256 e913cd82ee940195bf09e46835e0d59d4517a6ee3a3499c8556056e04883c0e1; original binding lines 189–189; characters 12367–12593; excerpt offset 0–226

Student work: Do not upload identifiable student work, grades, accommodations, or advising information unless the service is approved for the information and the use is consistent with FERPA and course expectations.

証拠の所在の検索: Captured source text; Teaching scenarios: Student work (research locator context; actual captured line range below); Extracted text SHA256 e913cd82ee940195bf09e46835e0d59d4517a6ee3a3499c8556056e04883c0e1; original binding lines 187–187; characters 12149–12364; excerpt offset 0–215

Teaching in the presence of ubiquitous GenAI tools should be considered a dynamic endeavor and necessitates flexibility. Approaches and expectations will vary by instructor, course, assessment, and activity. To ensure transparency, it is critical that expectations are clearly articulated in the syllabus and continually reinforced when assignments are given. Instructors should align their course policies with their college’s academic misconduct policies. Instructors are disciplinary experts responsible for what is taught, appropriate pedagogies, and assessment methodologies. GenAI may influence all three, and as such, instructors need flexibility to allow or disallow the use of GenAI tools. I

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 191da3aaa6036192d20101c7a9262c2edae1fcfad9788b08f28a5c08c16c567d; original binding lines 93–105; characters 939–3497; excerpt offset 0–700

f the latter policy is adopted, instructors should be aware that U-M does not currently support the use of surveillance and plagiarism detection tools as they cannot be reliably counted upon. Further, clear policies have to be formulated regarding punitive measures and how they may be enforced. If GenAI tools are allowed, the instructor should be clear which tools are allowed and in what capacity. It is also important to point out to students that although AI-generated content may appear reliable and factual, that is not always the case, and that inaccuracies, untruths, and false citations may be missed if one is overly reliant on the tools. Syllabus Language GenAI is changing rapidly, and n

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 191da3aaa6036192d20101c7a9262c2edae1fcfad9788b08f28a5c08c16c567d; original binding lines 93–105; characters 939–3497; excerpt offset 700–1400

ew tools will become available. Course policies therefore need to be provisional and subject to change. Several principles can guide instructors to develop course policies appropriate for their contexts: Protect cognitive dimension of learning: GenAI should enhance, not hinder, learning. Responsibility for content: Students are responsible for all content (ideas, facts, citations), however the work is generated. (Note: Chat GPT can generate untrue, biased, inaccurate, and hallucinatory content.) Require ethical transparency: Students should be transparent about how they use GenAI and adhere to standards of academic integrity. Explain academic misconduct: Instructors should distinguish betwee

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 191da3aaa6036192d20101c7a9262c2edae1fcfad9788b08f28a5c08c16c567d; original binding lines 93–105; characters 939–3497; excerpt offset 1400–2100

n allowed and disallowed uses of GenAI, including how, when, and why GenAI may (not) be used. Course policies might fall into one of three categories: Specific uses of GenAI are encouraged (generating ideas, editing, translating, outlining) Specific uses of GenAI are allowed if students clearly distinguish between their original work and GenAI output (highlighting output, tracking changes in GenAI output) Any use of GenAI constitutes academic misconduct.

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 191da3aaa6036192d20101c7a9262c2edae1fcfad9788b08f28a5c08c16c567d; original binding lines 93–105; characters 939–3497; excerpt offset 2100–2558

Use approved ITS AI Services.

証拠の所在の検索: Captured source text; When Handling University Data (research locator context; actual captured line range below); Extracted text SHA256 b8a1a567ae1ff37bf920e088aff945c8e324bf81e238d8a27c391c8a1a68396d; original binding lines 126–126; characters 2142–2171; excerpt offset 0–29

Only use university data classified as Low with AI services that do not have a contract or data agreement with U-M.

証拠の所在の検索: Captured source text; When Handling University Data (research locator context; actual captured line range below); Extracted text SHA256 b8a1a567ae1ff37bf920e088aff945c8e324bf81e238d8a27c391c8a1a68396d; original binding lines 127–127; characters 2172–2287; excerpt offset 0–115

Google Gemini can be used for certain kinds of sensitive data, as detailed in the Sensitive Data Guide to IT Services. You can access Gemini via the web app at https://gemini.google.com or mobile app (Android and iOS). Learn more about Google Gemini features, access and support resources. Gemini Notebook Google Gemini Notebook (formerly NotebookLM) is an AI-powered research and note-taking assistant available to all University of Michigan faculty, staff, and students. (Sponsored affiliates, alumni, and retirees do not have access to Gemini Notebook.) Using Google's Gemini models, Gemini Notebook lets you upload your own sources, such as documents, PDFs, and Google Docs, and interact with the

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 d51f3e5f38ee9862251ccbee5c32b2714c46642af4ac243ea096f572eec128d8; original binding lines 217–224; characters 3537–4751; excerpt offset 0–700

m through summaries, questions, audio overviews, flashcards, AI-assisted exploration, and more. It is especially helpful for synthesizing large amounts of information across uploaded materials. Google Gemini Notebook can be used for certain kinds of sensitive data, as detailed in the Sensitive Data Guide to IT Services. You can access Gemini Notebook via the web app at https://notebook.google.com or mobile app (Android and iOS).  Learn more about Google Gemini Notebook features, access, and support resources.

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 d51f3e5f38ee9862251ccbee5c32b2714c46642af4ac243ea096f572eec128d8; original binding lines 217–224; characters 3537–4751; excerpt offset 700–1214

ITS AI Services include the safeguards required by HIPAA. Accordingly, you may use these services with Protected Health Information (PHI). Check the ITS AI Service entry in the Sensitive Data Guide to IT Services for a list of the data types that are permitted to be used with ITS AI Services.

証拠の所在の検索: Captured source text; Sensitive Data with U-M AI Service Offerings (research locator context; actual captured line range below); Extracted text SHA256 4c61e5ace8770a2a884b98ccc67bb96753cb05222b39c1fe2104fe6c6a01cfa1; original binding lines 256–256; characters 5520–5813; excerpt offset 0–293

ITS AI Services include the safeguards required by HIPAA. Accordingly, you may use these services with Protected Health Information (PHI). Check the ITS AI Service entry in the Sensitive Data Guide to IT Services for a list of the data types that are permitted to be used with ITS AI Services.

証拠の所在の検索: Captured source text; Sensitive Data with U-M AI Service Offerings (research locator context; actual captured line range below); Extracted text SHA256 4c61e5ace8770a2a884b98ccc67bb96753cb05222b39c1fe2104fe6c6a01cfa1; original binding lines 256–256; characters 5520–5813; excerpt offset 0–293

Understand that in most cases, the data you share is not private and will be accessible by external parties hosting the GenAI-based tools. Do not share information that is considered private, or sensitive, such as credit card information, personal details such as ID numbers or addresses, and so on.

証拠の所在の検索: Captured source text; Choosing a GenAI tool to use: privacy risks (research locator context; actual captured line range below); Extracted text SHA256 409c6146855d44dd51c420ef1596ec12a6bde75cfbba018904934064b88a388b; original binding lines 100–100; characters 1863–2162; excerpt offset 0–299

Data shared with ITS AI Services by you will not be used to train the AI models made available by the U-M through the ITS AI Services.

証拠の所在の検索: Captured source text; At a Glance (research locator context; actual captured line range below); Extracted text SHA256 e4d24e7ad862879311fb40746d4d847389ea2d65ef2f1bb4606b504c8ae69a17; original binding lines 250–250; characters 3968–4102; excerpt offset 0–134

Data shared with ITS AI Services by you will not be used to train the AI models made available by the U-M through the ITS AI Services.

証拠の所在の検索: Captured source text; At a Glance (research locator context; actual captured line range below); Extracted text SHA256 e4d24e7ad862879311fb40746d4d847389ea2d65ef2f1bb4606b504c8ae69a17; original binding lines 250–250; characters 3968–4102; excerpt offset 0–134

Provide persistence of user prompt history across sessions with the ITS AI services.

証拠の所在の検索: Captured source text; How We Use Personal Information (research locator context; actual captured line range below); Extracted text SHA256 e4d24e7ad862879311fb40746d4d847389ea2d65ef2f1bb4606b504c8ae69a17; original binding lines 275–275; characters 6381–6465; excerpt offset 0–84

We do not sell or license your personal information collected here. To operate the ITS AI Services, however, we must share your prompt data with the cloud vendor utilized to host the ITS AI services. 

証拠の所在の検索: Captured source text; With Whom We Share Personal Information (research locator context; actual captured line range below); Extracted text SHA256 e4d24e7ad862879311fb40746d4d847389ea2d65ef2f1bb4606b504c8ae69a17; original binding lines 282–282; characters 7120–7320; excerpt offset 0–200

Crash reports and debugging logs

証拠の所在の検索: Captured source text; What Information We Collect (research locator context; actual captured line range below); Extracted text SHA256 e4d24e7ad862879311fb40746d4d847389ea2d65ef2f1bb4606b504c8ae69a17; original binding lines 269–269; characters 5996–6028; excerpt offset 0–32

What Information We Collect ITS AI Services collect your input into the AI applications when you directly provide it to us, such as when you log in, enter prompts or questions, and upload documents. ITS AI Services also capture some information about you automatically when you use the services. Automatically collected information may include: The Internet domain you use Your device’s IP address Your Internet Service Provider Your Cell Phone Service Provider The type of cell phone The type of browser The date and time of visit The address of the website that linked to ITS AI Services Information about your role at the U-M and other related HR and student data via Single Sign On. Crash reports

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 e4d24e7ad862879311fb40746d4d847389ea2d65ef2f1bb4606b504c8ae69a17; original binding lines 257–286; characters 5309–7958; excerpt offset 0–700

and debugging logs How We Use Personal Information The University of Michigan uses information about your interactions with ITS AI Services to: Monitor performance and make improvements to functionality and user experience. Provide support for questions about using the web applications or troubleshoot issues. Enforce token quota limitation and enable internal billing. Provide persistence of user prompt history across sessions with the ITS AI services. We use de-identified and/or aggregated information to: Monitor trends on the use of the tool to inform the product roadmap. Monitor application performance to inform improvements to the application, such as device and browser compatibility, ap

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 e4d24e7ad862879311fb40746d4d847389ea2d65ef2f1bb4606b504c8ae69a17; original binding lines 257–286; characters 5309–7958; excerpt offset 700–1400

plication features, etc. Develop projections for future usage and anticipate service costs. The use of personal information collected by this service is safeguarded through several mechanisms, including Information Assurance and U-M policies, such as Privacy and the Need to Monitor and Access Records (SPG 601.11) and Institutional Data Stewardship Policy (SPG 601.12). With Whom We Share Personal Information We do not sell or license your personal information collected here. To operate the ITS AI Services, however, we must share your prompt data with the cloud vendor utilized to host the ITS AI services.  Specifically, we share your information with Microsoft Azure Services and Google. Learn

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 e4d24e7ad862879311fb40746d4d847389ea2d65ef2f1bb4606b504c8ae69a17; original binding lines 257–286; characters 5309–7958; excerpt offset 1400–2100

more about Microsoft Azure’s privacy policy and Google's privacy policy. We require U-M’s service providers to keep your personal information secure and do not allow them to use or share it for any purpose other than providing services on our behalf.  In addition, you can choose to share your U-M Maizey experience with other U-M community members. We may also share your personal information when required by law, or when we believe sharing will help to protect the safety, property, or rights of U-M, members of the U-M community, and U-M guests.

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 e4d24e7ad862879311fb40746d4d847389ea2d65ef2f1bb4606b504c8ae69a17; original binding lines 257–286; characters 5309–7958; excerpt offset 2100–2649

AI services that will be used with institutional data classified as Moderate, High, or Restricted should only be purchased by Procurement Services.

証拠の所在の検索: Captured source text; Procurement of AI Services (research locator context; actual captured line range below); Extracted text SHA256 b8a1a567ae1ff37bf920e088aff945c8e324bf81e238d8a27c391c8a1a68396d; original binding lines 189–189; characters 4508–4655; excerpt offset 0–147

Google Gemini can be used for certain kinds of sensitive data, as detailed in the Sensitive Data Guide to IT Services. You can access Gemini via the web app at https://gemini.google.com or mobile app (Android and iOS). Learn more about Google Gemini features, access and support resources. Gemini Notebook Google Gemini Notebook (formerly NotebookLM) is an AI-powered research and note-taking assistant available to all University of Michigan faculty, staff, and students. (Sponsored affiliates, alumni, and retirees do not have access to Gemini Notebook.) Using Google's Gemini models, Gemini Notebook lets you upload your own sources, such as documents, PDFs, and Google Docs, and interact with the

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 d51f3e5f38ee9862251ccbee5c32b2714c46642af4ac243ea096f572eec128d8; original binding lines 217–224; characters 3537–4751; excerpt offset 0–700

m through summaries, questions, audio overviews, flashcards, AI-assisted exploration, and more. It is especially helpful for synthesizing large amounts of information across uploaded materials. Google Gemini Notebook can be used for certain kinds of sensitive data, as detailed in the Sensitive Data Guide to IT Services. You can access Gemini Notebook via the web app at https://notebook.google.com or mobile app (Android and iOS).  Learn more about Google Gemini Notebook features, access, and support resources.

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 d51f3e5f38ee9862251ccbee5c32b2714c46642af4ac243ea096f572eec128d8; original binding lines 217–224; characters 3537–4751; excerpt offset 700–1214

Ensure AI-generated computer code is always reviewed by a human.

証拠の所在の検索: Captured source text; To Generate Computer Code (research locator context; actual captured line range below); Extracted text SHA256 b8a1a567ae1ff37bf920e088aff945c8e324bf81e238d8a27c391c8a1a68396d; original binding lines 148–148; characters 2797–2861; excerpt offset 0–64

Conduct code reviews with professionally trained peers for all new or significantly changed applications, particularly those that maintain, process, transmit, or store data classified as Restricted or High.

証拠の所在の検索: Captured source text; To Generate Computer Code (research locator context; actual captured line range below); Extracted text SHA256 b8a1a567ae1ff37bf920e088aff945c8e324bf81e238d8a27c391c8a1a68396d; original binding lines 149–149; characters 2862–3068; excerpt offset 0–206

●​ The use of generative AI tools in peer review is discouraged. These tools struggle to assess novel and original ideas, may produce biased results, and have been shown to hallucinate. ●​ Secure AI platforms that are available through U-M ITS AI Services may complement the review process to address grammar, language, and writing clarity. If permitted by the funding agency (see below), certain AI tools may also be useful when analyzing data. ●​ Uploading, copying and pasting, or otherwise incorporating any part of another researcher’s grant application to a commercially available AI platform (e.g., chatGPT) is

証拠の所在の検索: Captured source text; PDF physical page 1; OVPR guidance bullet 1 (research locator context; actual captured line range below); Extracted text SHA256 5f306f37d8acd8d7bc85945eabd591c54518198b5ba8f983de56702118f2293c; original binding lines 11–18; characters 477–1137; excerpt offset 0–660

●​ Uploading, copying and pasting, or otherwise incorporating any part of another researcher’s grant application to a commercially available AI platform (e.g., chatGPT) is strongly discouraged. Such platforms lack sufficient privacy protections and may utilize submitted information, whether through prompts or otherwise, to train the models used in such services. This means that the submitted information could be disclosed through subsequent use of the platform by others. ●​ Please refer to the following guidelines to determine the appropriateness of using U-M ITS AI services: ○​ https://safecomputing.umich.edu/dataguide/service/75

証拠の所在の検索: Captured source text; PDF physical page 1; OVPR guidance bullet 3 (research locator context; actual captured line range below); Extracted text SHA256 5f306f37d8acd8d7bc85945eabd591c54518198b5ba8f983de56702118f2293c; original binding lines 17–25; characters 959–1654; excerpt offset 0–695

ITS is now offering a generative AI platform available to all active U-M faculty, staff, and students on the Ann Arbor, Flint, and Dearborn campuses and Michigan Medicine. These service offerings are equitable, accessible, and support everything from basic consumer usage to advanced research and experimentation.

証拠の所在の検索: Captured source text; ITS AI Services, platform access (research locator context; actual captured line range below); Extracted text SHA256 4c61e5ace8770a2a884b98ccc67bb96753cb05222b39c1fe2104fe6c6a01cfa1; original binding lines 213–213; characters 2906–3219; excerpt offset 0–313

Regardless of the physical location from which you access or use the Services, which shall be limited to use within the United States, you will be considered to be accessing or using information technology of U-M.

証拠の所在の検索: Captured source text; 1. Your Access (research locator context; actual captured line range below); Extracted text SHA256 8dc1654df028a8f4c8a7192877e79df266042264830147f6da2c8503a2177d67; original binding lines 219–219; characters 3513–3726; excerpt offset 0–213

Adobe Firefly is available for free through the University of Michigan's enterprise Adobe Creative Cloud license to all University of Michigan faculty, staff, and students. Firefly is a family of creative generative AI models used to generate, edit, and refine visual and audio content. It is designed specifically for commercial safety and is built directly into Adobe applications like Photoshop, Illustrator, and Premiere, as well as being accessible as a standalone web tool.

証拠の所在の検索: Captured source text; Other U-M AI Offerings (research locator context; actual captured line range below); Extracted text SHA256 c7010c47f8683762fc8ee63337cce319e7cfa766232915e9abf0e2e039d86eef; original binding lines 98–98; characters 2212–2691; excerpt offset 0–479

Regardless of the physical location from which you access or use the Services, which shall be limited to use within the United States, you will be considered to be accessing or using information technology of U-M.

証拠の所在の検索: Captured source text; 1. Your Access (research locator context; actual captured line range below); Extracted text SHA256 8dc1654df028a8f4c8a7192877e79df266042264830147f6da2c8503a2177d67; original binding lines 219–219; characters 3513–3726; excerpt offset 0–213

This list is for purely informational purposes. The University of Michigan does not endorse or support any of these specific external AI tools. There are many AI tools available and this is not an exhaustive list. 

証拠の所在の検索: Captured source text; AI Tools External to U-M (research locator context; actual captured line range below); Extracted text SHA256 c7010c47f8683762fc8ee63337cce319e7cfa766232915e9abf0e2e039d86eef; original binding lines 102–102; characters 3251–3465; excerpt offset 0–214

Google Gemini and Gemini Notebook are Google generative AI services available to all University of Michigan faculty, staff, and students. Google Gemini can be used to create new content and images, help brainstorm ideas, draft or summarize writing, explore complex topics, and assist with research and creative work. Gemini Notebook is an AI-powered research and note-taking assistant available to all University of Michigan faculty, staff, and students. Refer to the ITS Google Gemini & Gemini Notebook service page for more information on the U-M Google GenAI Services. 

証拠の所在の検索: Captured source text; U-M Google AI Tools (research locator context; actual captured line range below); Extracted text SHA256 c7010c47f8683762fc8ee63337cce319e7cfa766232915e9abf0e2e039d86eef; original binding lines 96–96; characters 1616–2188; excerpt offset 0–572

Google Gemini is a generative AI chat and image generation app available to all University of Michigan faculty, staff, and students. (Sponsored affiliates, alumni, and retirees do not have access to Gemini.) Like U-M’s GenAI services, Gemini can be used to create new content, generate and edit images (using Nano Banana), help brainstorm ideas, draft or summarize writing, explore complex topics, and assist with research and creative work. You can connect and use some Google Workspace services within the Gemini web app.

証拠の所在の検索: Captured source text; Google Gemini (research locator context; actual captured line range below); Extracted text SHA256 d51f3e5f38ee9862251ccbee5c32b2714c46642af4ac243ea096f572eec128d8; original binding lines 216–216; characters 3013–3536; excerpt offset 0–523

Regardless of the physical location from which you access or use the Services, which shall be limited to use within the United States, you will be considered to be accessing or using information technology of U-M.

証拠の所在の検索: Captured source text; 1. Your Access (research locator context; actual captured line range below); Extracted text SHA256 8dc1654df028a8f4c8a7192877e79df266042264830147f6da2c8503a2177d67; original binding lines 219–219; characters 3513–3726; excerpt offset 0–213

ITS is now offering a generative AI platform available to all active U-M faculty, staff, and students on the Ann Arbor, Flint, and Dearborn campuses and Michigan Medicine. These service offerings are equitable, accessible, and support everything from basic consumer usage to advanced research and experimentation.

証拠の所在の検索: Captured source text; ITS AI Services, platform access (research locator context; actual captured line range below); Extracted text SHA256 4c61e5ace8770a2a884b98ccc67bb96753cb05222b39c1fe2104fe6c6a01cfa1; original binding lines 213–213; characters 2906–3219; excerpt offset 0–313

ITS offers a generative AI platform available to all active U-M faculty, staff, and students on the Ann Arbor, Flint, and Dearborn campuses and Michigan Medicine. These service offerings, which include U-M GPT, U-M Maizey, and U-M GPT Toolkit, are secure, private, equitable, accessible, and free. They support everything from basic consumer usage to advanced research and experimentation.

証拠の所在の検索: Captured source text; U-M AI Tools (research locator context; actual captured line range below); Extracted text SHA256 c7010c47f8683762fc8ee63337cce319e7cfa766232915e9abf0e2e039d86eef; original binding lines 94–94; characters 1206–1595; excerpt offset 0–389

Regardless of the physical location from which you access or use the Services, which shall be limited to use within the United States, you will be considered to be accessing or using information technology of U-M.

証拠の所在の検索: Captured source text; 1. Your Access (research locator context; actual captured line range below); Extracted text SHA256 8dc1654df028a8f4c8a7192877e79df266042264830147f6da2c8503a2177d67; original binding lines 219–219; characters 3513–3726; excerpt offset 0–213

Our most accessible ITS AI service is U-M GPT, a conversational AI tool that provides access to popular hosted AI models such as Azure OpenAI, GPT-Image-1, and U-M hosted open-source large language models.

証拠の所在の検索: Captured service-description paragraph; actual line range recorded; Extracted text SHA256 4c61e5ace8770a2a884b98ccc67bb96753cb05222b39c1fe2104fe6c6a01cfa1; original binding lines 215–215; characters 3228–3433; excerpt offset 0–205

ITS is now offering a generative AI platform available to all active U-M faculty, staff, and students on the Ann Arbor, Flint, and Dearborn campuses and Michigan Medicine. These service offerings are equitable, accessible, and support everything from basic consumer usage to advanced research and experimentation.

証拠の所在の検索: Captured source text; ITS AI Services, platform access (research locator context; actual captured line range below); Extracted text SHA256 4c61e5ace8770a2a884b98ccc67bb96753cb05222b39c1fe2104fe6c6a01cfa1; original binding lines 213–213; characters 2906–3219; excerpt offset 0–313

ITS offers a generative AI platform available to all active U-M faculty, staff, and students on the Ann Arbor, Flint, and Dearborn campuses and Michigan Medicine. These service offerings, which include U-M GPT, U-M Maizey, and U-M GPT Toolkit, are secure, private, equitable, accessible, and free. They support everything from basic consumer usage to advanced research and experimentation.

証拠の所在の検索: Captured source text; U-M AI Tools (research locator context; actual captured line range below); Extracted text SHA256 c7010c47f8683762fc8ee63337cce319e7cfa766232915e9abf0e2e039d86eef; original binding lines 94–94; characters 1206–1595; excerpt offset 0–389

Regardless of the physical location from which you access or use the Services, which shall be limited to use within the United States, you will be considered to be accessing or using information technology of U-M.

証拠の所在の検索: Captured source text; 1. Your Access (research locator context; actual captured line range below); Extracted text SHA256 8dc1654df028a8f4c8a7192877e79df266042264830147f6da2c8503a2177d67; original binding lines 219–219; characters 3513–3726; excerpt offset 0–213

U-M GPT Toolkit is our most advanced and flexible service offering. It is designed for those who require full control over their AI environments and models. Access to the API gateway is available via the self-service portal.

証拠の所在の検索: Captured service-description paragraph; actual line range recorded; Extracted text SHA256 4c61e5ace8770a2a884b98ccc67bb96753cb05222b39c1fe2104fe6c6a01cfa1; original binding lines 219–219; characters 3732–3956; excerpt offset 0–224

ITS is now offering a generative AI platform available to all active U-M faculty, staff, and students on the Ann Arbor, Flint, and Dearborn campuses and Michigan Medicine. These service offerings are equitable, accessible, and support everything from basic consumer usage to advanced research and experimentation.

証拠の所在の検索: Captured source text; ITS AI Services, platform access (research locator context; actual captured line range below); Extracted text SHA256 4c61e5ace8770a2a884b98ccc67bb96753cb05222b39c1fe2104fe6c6a01cfa1; original binding lines 213–213; characters 2906–3219; excerpt offset 0–313

ITS offers a generative AI platform available to all active U-M faculty, staff, and students on the Ann Arbor, Flint, and Dearborn campuses and Michigan Medicine. These service offerings, which include U-M GPT, U-M Maizey, and U-M GPT Toolkit, are secure, private, equitable, accessible, and free. They support everything from basic consumer usage to advanced research and experimentation.

証拠の所在の検索: Captured source text; U-M AI Tools (research locator context; actual captured line range below); Extracted text SHA256 c7010c47f8683762fc8ee63337cce319e7cfa766232915e9abf0e2e039d86eef; original binding lines 94–94; characters 1206–1595; excerpt offset 0–389

Regardless of the physical location from which you access or use the Services, which shall be limited to use within the United States, you will be considered to be accessing or using information technology of U-M.

証拠の所在の検索: Captured source text; 1. Your Access (research locator context; actual captured line range below); Extracted text SHA256 8dc1654df028a8f4c8a7192877e79df266042264830147f6da2c8503a2177d67; original binding lines 219–219; characters 3513–3726; excerpt offset 0–213

U-M Maizey is a tool that allows U-M faculty, staff, and students to enrich their GenAI experience based on a custom dataset they provide. This service empowers users to extract valuable insights, discover patterns, and gain deeper knowledge from the available datasets.

証拠の所在の検索: Captured service-description paragraph; actual line range recorded; Extracted text SHA256 4c61e5ace8770a2a884b98ccc67bb96753cb05222b39c1fe2104fe6c6a01cfa1; original binding lines 217–217; characters 3445–3715; excerpt offset 0–270

Regardless of the physical location from which you access or use the Services, which shall be limited to use within the United States, you will be considered to be accessing or using information technology of U-M.

証拠の所在の検索: Captured source text; 1. Your Access (research locator context; actual captured line range below); Extracted text SHA256 8dc1654df028a8f4c8a7192877e79df266042264830147f6da2c8503a2177d67; original binding lines 219–219; characters 3513–3726; excerpt offset 0–213

Google Gemini is a generative AI chat and image generation app available to all University of Michigan faculty, staff, and students. (Sponsored affiliates, alumni, and retirees do not have access to Gemini.) Like U-M’s GenAI services, Gemini can be used to create new content, generate and edit images (using Nano Banana), help brainstorm ideas, draft or summarize writing, explore complex topics, and assist with research and creative work. You can connect and use some Google Workspace services within the Gemini web app.

証拠の所在の検索: Captured source text; Google Gemini (research locator context; actual captured line range below); Extracted text SHA256 d51f3e5f38ee9862251ccbee5c32b2714c46642af4ac243ea096f572eec128d8; original binding lines 216–216; characters 3013–3536; excerpt offset 0–523

Google Gemini Notebook (formerly NotebookLM) is an AI-powered research and note-taking assistant available to all University of Michigan faculty, staff, and students. (Sponsored affiliates, alumni, and retirees do not have access to Gemini Notebook.) Using Google's Gemini models, Gemini Notebook lets you upload your own sources, such as documents, PDFs, and Google Docs, and interact with them through summaries, questions, audio overviews, flashcards, AI-assisted exploration, and more. It is especially helpful for synthesizing large amounts of information across uploaded materials.

証拠の所在の検索: Captured source text; Gemini Notebook (research locator context; actual captured line range below); Extracted text SHA256 d51f3e5f38ee9862251ccbee5c32b2714c46642af4ac243ea096f572eec128d8; original binding lines 221–221; characters 3843–4430; excerpt offset 0–587

Google Gemini can be used for certain kinds of sensitive data, as detailed in the Sensitive Data Guide to IT Services. You can access Gemini via the web app at https://gemini.google.com or mobile app (Android and iOS). Learn more about Google Gemini features, access and support resources. Gemini Notebook Google Gemini Notebook (formerly NotebookLM) is an AI-powered research and note-taking assistant available to all University of Michigan faculty, staff, and students. (Sponsored affiliates, alumni, and retirees do not have access to Gemini Notebook.) Using Google's Gemini models, Gemini Notebook lets you upload your own sources, such as documents, PDFs, and Google Docs, and interact with the

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 d51f3e5f38ee9862251ccbee5c32b2714c46642af4ac243ea096f572eec128d8; original binding lines 217–224; characters 3537–4751; excerpt offset 0–700

m through summaries, questions, audio overviews, flashcards, AI-assisted exploration, and more. It is especially helpful for synthesizing large amounts of information across uploaded materials. Google Gemini Notebook can be used for certain kinds of sensitive data, as detailed in the Sensitive Data Guide to IT Services. You can access Gemini Notebook via the web app at https://notebook.google.com or mobile app (Android and iOS).  Learn more about Google Gemini Notebook features, access, and support resources.

証拠の所在の検索: Captured source text; bounded full context, actual line range recorded; Extracted text SHA256 d51f3e5f38ee9862251ccbee5c32b2714c46642af4ac243ea096f572eec128d8; original binding lines 217–224; characters 3537–4751; excerpt offset 700–1214