University of Pennsylvania

In all cases, you should not upload the student’s work into an “AI detector.” Uploading a student’s work without their consent is a violation of FERPA and Penn’s privacy policies. These tools are also not accurate enough to be treated as evidence of misconduct as they exhibit frequent bias.

Bevisfinner: FAQ > suspected inappropriate AI use; character chunk 0; extracted text lines 366–373

Penn does not have a standardized policy on student AI use, and each instructor is responsible for setting their own guidelines. Your department or program may have recommendations or resources for developing your policies in addition to the considerations provided here.

Bevisfinner: Creating AI Policies > introduction; character chunk 0; extracted text lines 204–206

Individual instructors determine their own policies related to acceptable student use of generative AI in coursework. For more information about setting expectations for student AI use, visit the Course Policies & Expectations page.

Bevisfinner: Additional Guidance on AI Use; character chunk 0; extracted text lines 284–286

In the absence of other guidance, treat the use of AI as you would treat assistance from another person.  For example, this means if it is unacceptable to have another person substantially complete a task like writing an essay, it is also unacceptable to have AI to complete the task.

Bevisfinner: In the absence of other guidance,; character chunk 0; extracted text lines 103–103

At the discretion of faculty, instructional teams and staff, Wharton Academy students may use generative AI tools. Please note, however, that the material generated by these tools may produce inaccurate, incomplete, or otherwise problematic content. While these tools, if used properly, can generate new ideas, using them can also stifle students’ independent thinking and creativity. When students express themselves in Wharton Academy, they must use their voice and words. Using somebody else’s work without crediting the source – including generative AI — is plagiarism. Guided by the policies of faculty, instructional teams, and staff, AI-generated work should be cited like

Bevisfinner: Wharton Academy > Statement of AI Use; character chunk 0; extracted text lines 36–40

any other reference material, including how and where students used AI-generated information. While AI can be a helpful tool, students should be mindful of its limitations and ensure their final work reflects their ideas, words, and creativity. As students navigate the emerging AI landscape, they can look to the Wharton School of the University of Pennsylvania as a world leader in AI and a reliable source of AI insight and guidance. For more details on academic honesty, please refer to Penn’s Academic Integrity statement.

Bevisfinner: Wharton Academy > Statement of AI Use; character chunk 680; extracted text lines 40–49

DON'TS Don’t paste your homework/exam questions directly as prompts. Irresponsible prompt: “Solve this: 3x + 5 = 14” Irresponsible prompt: “Write a 5-paragraph essay on the causes of World War II.” Don’t directly copy answers from generative AI tools and submit as your own. Don’t use AI for personal reflection or opinion-based tasks. We are interested in hearing your opinions, your stories and your thoughts. In our classrooms, your fellow students hope to learn from you, your background, your understanding on the course materials and your thought process, not from a machine. Don’t rely on AI for group assignments. Using AI to complete your portion of a group project inste

Bevisfinner: Wharton Academy > Don'ts; character chunk 0; extracted text lines 82–94

ad of collaborating with your peers is considered academic dishonesty. Your peers hope to work with you as a team member, not with a machine. Don’t use AI to cheat on exams or tests. Don’t use AI to paraphrase or rewrite plagiarized content. Don’t use AI to fake research or sources. Don’t post AI-generated discussion posts within the course community forum.

Bevisfinner: Wharton Academy > Don'ts; character chunk 680; extracted text lines 94–98

DON'TS Don’t paste your homework/exam questions directly as prompts. Irresponsible prompt: “Solve this: 3x + 5 = 14” Irresponsible prompt: “Write a 5-paragraph essay on the causes of World War II.” Don’t directly copy answers from generative AI tools and submit as your own. Don’t use AI for personal reflection or opinion-based tasks. We are interested in hearing your opinions, your stories and your thoughts. In our classrooms, your fellow students hope to learn from you, your background, your understanding on the course materials and your thought process, not from a machine. Don’t rely on AI for group assignments. Using AI to complete your portion of a group project inste

Bevisfinner: Wharton Academy > Don'ts; character chunk 0; extracted text lines 82–94

ad of collaborating with your peers is considered academic dishonesty. Your peers hope to work with you as a team member, not with a machine. Don’t use AI to cheat on exams or tests. Don’t use AI to paraphrase or rewrite plagiarized content. Don’t use AI to fake research or sources. Don’t post AI-generated discussion posts within the course community forum.

Bevisfinner: Wharton Academy > Don'ts; character chunk 680; extracted text lines 94–98

Guidance for Educators Use of AI should always be in alignment with Penn’s Principles of Responsible Conduct . As expectations may vary between classes and instructors, it is important for instructors to provide students with clear guidelines similar to the guidelines on collaboration , on the use of AI within coursework, and when and how the use of AI within a course should be cited. Disclose to students when course materials have been created with the use of AI and when AI detection software will be used in the course. Be aware of how AI tools may impact assignments and exams in your discipline. Penn's Center for Excellence in Teaching, Learning and Innovation (CETLI) p

Bevisfinner: Guidance for Educators Use; character chunk 0; extracted text lines 84–92

rovides useful guidance in this space.

Bevisfinner: Guidance for Educators Use; character chunk 680; extracted text lines 92–94

In other cases, student AI use may not be as clear. Often, faculty suspect a student has used AI if the language in their writing is bland or vague, but sometimes those qualities reflect the student’s own work. If you suspect a student has used AI in a way that violates your course policy, have a conversation with the student. You can begin the conversation with questions or statements like: I’m interested in understanding how you reached this conclusion. Can you talk to me about the process you used to draft this? How did you think about the organization of your work? You may discuss suspected student misuse of AI with the Center for Community Standards & Accountability

Bevisfinner: FAQ > suspected inappropriate AI use; conversation and CSA consultation; character chunk 0; extracted text lines 359–365

, which handles academic integrity violations. Their staff can consult with you about your concerns and help you plan next steps, whether or not you intend to refer the student for further investigation.

Bevisfinner: FAQ > suspected inappropriate AI use; conversation and CSA consultation; character chunk 680; extracted text lines 365–365

Most AI tools and services use input and data from users of the tool to train the model.  Additionally, existing tools may incorporate AI features in their service offerings. For this reason, users of AI should avoid sharing personal or sensitive data with the tool and should not  input moderate or high-risk Penn data as defined by the Penn Data Risk Classification , or intellectual property, without: Careful consideration and understanding of the tool’s use of Penn data and the service provider’s stated rights to the data, including, but not limited to whether the service provider offers the option to opt-out of using customer’s data to train the AI; A contract in place

Bevisfinner: Privacy & Contracts; distinct data/IP consultation routes; character chunk 0; extracted text lines 64–68

to protect Penn data; and Review by Penn’s Privacy Office and consultation with the Office of Information Security as coordinated by Procurement when moderate or high-risk data is involved. Consultation with the Penn Center for Innovation, where intellectual property is involved.

Bevisfinner: Privacy & Contracts; distinct data/IP consultation routes; character chunk 680; extracted text lines 68–70

Patient Privacy Protection It is not permissible under the Health Information Portability & Accountability Act (HIPAA) or Penn Medicine policy to share patient or research participant information in connection with open or public AI tools and services, such as ChatGPT. This is because, as currently configured, such open or public tools and services can use and share any data without regard to HIPAA restrictions and other protections. Therefore, individual patient data and patient data sets (even if deidentified) may not be exposed to open or public AI tools or services, absent institutional approval.

Bevisfinner: Patient Privacy Protection; character chunk 0; extracted text lines 71–72

PennChat Free for Penn community •FERPA-protected educational records •HIPAA-protected health information Log in with PennKey while on campus or connected to VPN

Bevisfinner: Penn-Licensed AI Tools > PennChat row; character chunk 0; extracted text lines 257–261

While these Penn-licensed tools provide some added privacy protections, it is important to be aware of which tools are acceptable to use for different types of student data.

Bevisfinner: Licensed tool data scope; character chunk 0; extracted text lines 212–212

Do not require students to enter personal information or intellectual property into an unlicensed AI tool. Unlicensed tools may be used optionally by students at the instructor’s discretion but consider using a Penn-licensed tool for mandatory components of coursework to protect student data.

Bevisfinner: Unlicensed tools; character chunk 0; extracted text lines 280–281

Do not enter any information that could identify a student. This includes names, ID numbers, or email addresses, as well as detailed descriptions of student work or engagement in class that could be identifiable to others.

Bevisfinner: Publicly Available AI Tools; character chunk 0; extracted text lines 274–275

Do not enter student work (e.g., papers, projects) without the student’s permission, even if it is anonymized . This work is part of the student’s confidential academic record.

Bevisfinner: Publicly Available AI Tools; character chunk 0; extracted text lines 276–279

Do not require students to enter personal information or intellectual property into an unlicensed AI tool. Unlicensed tools may be used optionally by students at the instructor’s discretion but consider using a Penn-licensed tool for mandatory components of coursework to protect student data.

Bevisfinner: Publicly Available AI Tools; character chunk 0; extracted text lines 280–281

Intellectual Property Members of the Penn community should adhere to established principles of respect for intellectual property, particularly copyrights when considering the creation of new data sets for training AI models.  Avoid uploading confidential and/or proprietary information to AI platforms prior to seeking patent or copyright protection, as doing so could jeopardize IP rights.

Bevisfinner: Intellectual Property Members; character chunk 0; extracted text lines 79–80

Researchers should adhere to federal or international requirements on obtaining informed consent, and Institutional Review Board approvals should be obtained prior to exposing research participant data to AI tools. Caution should be adopted when research involves the examination of high-risk data, including Personally Identifiable Information (PII) and research participant health information (both identifiable and non-identifiable) exposed to AI.

Bevisfinner: Researchers should adhere; character chunk 0; extracted text lines 107–108

Consult with your department leadership and your discipline’s publishing standards to determine how the use of AI should be accounted for with regard to authorship in publications.

Bevisfinner: Consult with your department leadership; character chunk 0; extracted text lines 106–106

ChatGPT Edu Full-range generative AI chat client (text, images, audio). FAQs can be found here. Available for faculty and staff at $13 per month; please contact your IT Support Provider . Students may have access when required in course or research experience, and paid by the School, department, or instructor; annual commitment required Low and Moderate Risk Data . Do not input sensitive data including HIPAA data, personal health information (PHI), financial account information, SSNs, and credit card data. Additional Guidance.

Bevisfinner: ISC tools table > ChatGPT Edu; character chunk 0; extracted text lines 127–136

ISC manages a range of Generative AI tools for campus with negotiated enterprise contracts including improved security, privacy, and cost. The following table provides information on available tools, costs, and data protections. For all of these tools, users must be logged into a University account for appropriate protections. Personal accounts on these services are not protected.

Bevisfinner: University-account protection condition; character chunk 0; extracted text lines 54–55

Gemini for Google Workspace Full-Range; business tools integrated Integrates generative AI into the Google Workspace. The tool is designed to streamline routine tasks and enhance productivity. Available for some faculty and staff at $20 or $30 per month only at the following schools: ASC, SEAS, and Wharton; please contact your IT Support Provider . Low and Moderate Risk Data. Do not input sensitive data including HIPAA data, personal health information (PHI), financial account information, SSNs, and credit card data. Additional Guidance .

Bevisfinner: ISC tools table > Gemini for Google Workspace; character chunk 0; extracted text lines 107–116

ISC manages a range of Generative AI tools for campus with negotiated enterprise contracts including improved security, privacy, and cost. The following table provides information on available tools, costs, and data protections. For all of these tools, users must be logged into a University account for appropriate protections. Personal accounts on these services are not protected.

Bevisfinner: University-account protection condition; character chunk 0; extracted text lines 54–55

Grammarly Pro Purchase required •FERPA-protected educational records •HIPAA-protected health information Contact local IT support provider

Bevisfinner: Penn-Licensed AI Tools > Grammarly Pro; character chunk 0; extracted text lines 247–251

PennChat PennChat is a generative AI chat client that runs on OpenAI's and Anthropic’s large language models to create textual content. It can also ingest, analyze, and output various document formats. Free to Penn community Low, Moderate, and most High Risk Data , except SSNs and credit card data.

Bevisfinner: ISC tools table > PennChat; character chunk 0; extracted text lines 63–67

ISC manages a range of Generative AI tools for campus with negotiated enterprise contracts including improved security, privacy, and cost. The following table provides information on available tools, costs, and data protections. For all of these tools, users must be logged into a University account for appropriate protections. Personal accounts on these services are not protected.

Bevisfinner: University-account protection condition; character chunk 0; extracted text lines 54–55

PennChat Free for Penn community •FERPA-protected educational records •HIPAA-protected health information Log in with PennKey while on campus or connected to VPN

Bevisfinner: CETLI > PennChat access/data row; character chunk 0; extracted text lines 257–261

Snowflake Data Analytics Platform & Consulting Full data, analytics, and AI platform using industry-leading LLMs from OpenAI, Anthropic, Meta, and others. Costs are based on consumption, not usage. Please contact your ISC Support Provider . Low, Moderate, and most High Risk Data , except HIPAA-protected, SSNs and credit card data.

Bevisfinner: ISC tools table > Snowflake Data Analytics Platform & Consulting; character chunk 0; extracted text lines 99–106

ISC manages a range of Generative AI tools for campus with negotiated enterprise contracts including improved security, privacy, and cost. The following table provides information on available tools, costs, and data protections. For all of these tools, users must be logged into a University account for appropriate protections. Personal accounts on these services are not protected.

Bevisfinner: University-account protection condition; character chunk 0; extracted text lines 54–55

Microsoft Copilot Chat (Basic)* Full-range generative AI chat client, running on ChatGPT, that provides web search enhanced with generative AI to create novel content (text, images). FAQs can be found here. No additional charge over existing Microsoft 365 agreements*. Copilot Chat is self-service and available at this site . Low, Moderate, and most High Risk Data , except SSNs and credit card data. Additional Guidance .

Bevisfinner: ISC tools table > Microsoft Copilot Chat (Basic); character chunk 0; extracted text lines 78–88

ISC manages a range of Generative AI tools for campus with negotiated enterprise contracts including improved security, privacy, and cost. The following table provides information on available tools, costs, and data protections. For all of these tools, users must be logged into a University account for appropriate protections. Personal accounts on these services are not protected.

Bevisfinner: University-account protection condition; character chunk 0; extracted text lines 54–55

Microsoft M365 Copilot (Premium)* Full-Range; business tools integrated Integrates OpenAI ChatGPT or Anthropic Claude into the Microsoft Office applications—Excel, OneNote, Outlook, PowerPoint, Teams, and Word. It also includes a Work search that allows queries across all of those applications. Available for faculty and staff at $16.50 per month; annual commitment required; please contact your IT Support Provider . Low, Moderate, and most High Risk Data , except SSNs, and credit card data. Additional Guidance .

Bevisfinner: ISC tools table > Microsoft M365 Copilot (Premium); character chunk 0; extracted text lines 89–98

ISC manages a range of Generative AI tools for campus with negotiated enterprise contracts including improved security, privacy, and cost. The following table provides information on available tools, costs, and data protections. For all of these tools, users must be logged into a University account for appropriate protections. Personal accounts on these services are not protected.

Bevisfinner: University-account protection condition; character chunk 0; extracted text lines 54–55

Google Gemini Notebook AI Research Assistant A research and note-taking assistant that focuses responses solely on your uploaded content, greatly reducing the risk of errors  (‘hallucinations’).  It can output text and audio. Available for some faculty and staff only at the following schools: ASC, SEAS, and Wharton; please contact your IT Support Provider . Low and Moderate Risk Data. Do not input sensitive data including HIPAA data, personal health information (PHI), financial account information, SSNs, and credit card data. Additional Guidance .

Bevisfinner: ISC tools table > Google Gemini Notebook; character chunk 0; extracted text lines 117–126

ISC manages a range of Generative AI tools for campus with negotiated enterprise contracts including improved security, privacy, and cost. The following table provides information on available tools, costs, and data protections. For all of these tools, users must be logged into a University account for appropriate protections. Personal accounts on these services are not protected.

Bevisfinner: University-account protection condition; character chunk 0; extracted text lines 54–55

Claude Anthropic Full-range generative AI chat client that provides web search enhanced with generative AI to create novel content (text, images, code). $16/user/month standard; $50/user/month premium annual commitment required Low and Moderate Risk Data. Do not input sensitive data including HIPAA data, personal health information (PHI), financial account information, SSNs, and credit card data. Additional Guidance .

Bevisfinner: ISC tools table > Claude Anthropic; character chunk 0; extracted text lines 68–77

ISC manages a range of Generative AI tools for campus with negotiated enterprise contracts including improved security, privacy, and cost. The following table provides information on available tools, costs, and data protections. For all of these tools, users must be logged into a University account for appropriate protections. Personal accounts on these services are not protected.

Bevisfinner: University-account protection condition; character chunk 0; extracted text lines 54–55

Adobe Firefly Free for Penn Community •FERPA-protected educational records Log in with PennKey

Bevisfinner: Penn-Licensed AI Tools > Adobe Firefly; character chunk 0; extracted text lines 217–220

OpenEvidence Available to PSOM users only •FERPA-protected educational records •HIPAA-protected health information Log in with Penn Medicine credentials

Bevisfinner: Penn-Licensed AI Tools > OpenEvidence; character chunk 0; extracted text lines 252–256

Zoom AI Companion Free for Penn community •FERPA-protected educational records Enable AI Companion features in your Zoom account settings

Bevisfinner: Penn-Licensed AI Tools > Zoom AI Companion; character chunk 0; extracted text lines 266–269

Which AI tools do members of the Penn Community have access to? Penn provides faculty, staff, and student access to various generative AI platforms. Some tools are available for free by logging in with your PennKey , while others require a budget code to purchase a user license. The list of provided tools changes frequently . Visit Penn AI Tools & Policies for available tools and instructions to access them. Do I have to use AI in my teaching? It’s up to individual instructors to determine if AI aligns with their teaching aims and if it’s something they see value in allowing or implementing within their courses. Some Penn instructors prohibit AI entirely or only allow it

Bevisfinner: FAQ > tools and instructor discretion; character chunk 0; extracted text lines 266–277

for very specific uses. Whether you choose to use AI in your teaching or not, you should communicate your expectations clearly with students .

Bevisfinner: FAQ > tools and instructor discretion; character chunk 680; extracted text lines 277–281

Be transparent about the use of AI.  Disclose when a work product was created wholly or partially using an AI tool and, if appropriate, how AI was used to create the work product.

Bevisfinner: General Guidance > Transparency; character chunk 0; extracted text lines 58–58

This statement is not intended as legal advice or an exhaustive set of best practices and should not be viewed as a final policy. The AI field is rapidly evolving in terms of technology, deployment models, third-party relationships, terms of service, regulatory landscape, and academic-industry partnership structures. It is anticipated that this document will be updated regularly and interact with other sources of policy, ethics, and governing legal authority.

Bevisfinner: Scope; character chunk 0; extracted text lines 55–55

Educators may have requirements and guidance for citing the use of generative AI output and for attributing AI created content to the specific AI tool and parameters used. Individual courses may have different or more narrow guidance on the use of AI that should be adhered to within the context of that course.

Bevisfinner: Guidance for Students; character chunk 0; extracted text lines 101–102

This information is subject to frequent changes and was last updated on October 7, 2026. For the most up-to-date information, including costs and instructions for access to AI tools, visit the Information Systems & Computing (ISC) page on Generative AI Tools & Resources .

Bevisfinner: Update notice dated October 7, 2026; character chunk 0; extracted text lines 204–207

Penn offers access to a number of AI tools. The guidelines for protecting student privacy  while using these tools are informed by the data risk classification and the privacy agreements of the tool being used. While these Penn-licensed tools provide some added privacy protections, it is important to be aware of which tools are acceptable to use for different types of student data.

Bevisfinner: Penn-Licensed AI Tools; privacy context; character chunk 0; extracted text lines 209–212

Individual courses may have different or more narrow guidance on the use of AI that should be adhered to within the context of that course.

Bevisfinner: Course-specific rules; character chunk 0; extracted text lines 102–102

Accountability The user of AI should endeavor to validate the accuracy of created content with trusted first party sources and monitor the reliability of that content. Users are accountable for their use of content created by AI and should be wary of misinformation or “hallucinations” by AI tools (e.g., citations to publications or source materials that do not exist or references that otherwise distort the truth).

Bevisfinner: Accountability; character chunk 0; extracted text lines 59–60

University Business Processes While automating tasks using AI may improve operational efficiency for University Business processes, oversight and review of the use of AI and verification of its outputs for these University business processes should be in place to ensure reliability, consistency, and accuracy.

Bevisfinner: University Business Processes While; character chunk 0; extracted text lines 81–82

At the discretion of faculty, instructional teams and staff, Wharton Academy students may use generative AI tools. Please note, however, that the material generated by these tools may produce inaccurate, incomplete, or otherwise problematic content. While these tools, if used properly, can generate new ideas, using them can also stifle students’ independent thinking and creativity. When students express themselves in Wharton Academy, they must use their voice and words. Using somebody else’s work without crediting the source – including generative AI — is plagiarism. Guided by the policies of faculty, instructional teams, and staff, AI-generated work should be cited like

Bevisfinner: Wharton Academy > Statement of AI Use; character chunk 0; extracted text lines 36–40

any other reference material, including how and where students used AI-generated information. While AI can be a helpful tool, students should be mindful of its limitations and ensure their final work reflects their ideas, words, and creativity. As students navigate the emerging AI landscape, they can look to the Wharton School of the University of Pennsylvania as a world leader in AI and a reliable source of AI insight and guidance. For more details on academic honesty, please refer to Penn’s Academic Integrity statement.

Bevisfinner: Wharton Academy > Statement of AI Use; character chunk 680; extracted text lines 40–49

DO'S Use it to help you brainstorm or boost your learning efficiency. Example prompt: “Can you organize my notes [include your notes here] into an outline?” Example prompt: “Can you create flashcards from these glossary terms [input a list of definitions you have learned in class]?” * Caution: Sharing your original work with generative AI tools may raise privacy concerns, as it is an AI tool that processes input to generate responses. Although user inputs are not directly integrated into the large language model, they are accessible to and stored by OpenAI and may be utilized to improve future versions of the tool. Ask for prompts. Example prompt: “Can you provide practic

Bevisfinner: Wharton Academy > Do's; instructor permission prerequisite; character chunk 0; extracted text lines 62–69

e problems for [the topic you want to practice more]?” * Caution: Remember the calculation limitation that generative AI tools have. Get both sides of the argument. Example prompt: “What are the different perspectives on X?” *Caution: Remember the limitation as an outdated data source if you want to gather information on more recent events. Also remember the limitation of fake citations. You should still do your own research to verify the different voices. Ask for templates. This is comparable to a traditional Google Search. Example prompt: “Can you provide me with a 1-page cover letter template?” Example prompt: “Can you show me a good infographic template?” Ask for feed

Bevisfinner: Wharton Academy > Do's; instructor permission prerequisite; character chunk 680; extracted text lines 69–77

back. You can ask for preliminary feedback on written work, which could sometimes be useful. *Caution: Remember the privacy concerns. Use it for language translation. Generative AI tools are powerful in providing more context-based translations. Example prompt: “Can you translate [a sentence] in the context of an [specific occasion] into [target language]?” *Caution: Remember the limitation as an outdated data source. Language transforms quickly and the answer you get might not still reflect how current speakers communicate.

Bevisfinner: Wharton Academy > Do's; instructor permission prerequisite; character chunk 1360; extracted text lines 77–81