You are here: American University Information Technology IT Security & Privacy University Guidance on the Responsible Use of Artificial Intelligence (AI)

Back to top

University Guidance on the Responsible Use of Artificial Intelligence (AI)

With the increasing integration of artificial intelligence (AI) tools—such as ChatGPT, Copilot, Perplexity, and other applications known as large language models (LLM) or generative AI applications—into university activities and existing platforms, it is essential to use these technologies responsibly. The following guidance outlines acceptable practices for utilizing AI tools while safeguarding institutional, personal, and proprietary information. This guidance will be updated over time to reflect changes in the regulatory environment, industry standards, and technology.

General Applicability

Principled Use

  • American University’s approach to Artificial Intelligence reflects our core values, including commitments to human dignity, equity, integrity, and free inquiry supporting truth-seeking, community, and positively impacting the world.  [See Principles document for more information]
  • Be transparent about the use of AI. Disclose if a work product was wholly or partially created using an AI tool and, if appropriate, how AI was used to create the work product. Considering context is appropriate, transparency may not look the same in all situations; the key is honesty about use.
  • The AI user is responsible for validating the accuracy of created content with trusted first-party sources and monitoring the reliability of that content. Users are accountable for their use of content created by AI and the impact of that use on their work.

No decisions without humans

  • All outcomes generated by AI must be subject to human oversight. Decisions impacting real persons in the American University community (e.g. employment, admissions, students) require human oversight. Members of the AU community are accountable for the work product or decisions resulting from their use of AI.

AI Literacy

  • Community members should develop foundational AI literacy by understanding AI fundamentals, critically evaluating AI applications in teaching and scholarship, and maintaining vigilance against bias and misuse. They should also commit to use in conformity with university mission and values, ensuring AI tools are applied in a responsible way with awareness of their impacts.

Data usage

  • Consider any information given to any open (non-enterprise) AI tool as if it were public.
  • Do not enter any sensitive data, or Confidential or Official Use data (as defined in the university Data Classification Policy) in an AI platform, unless approved by a university official. The Data Classification Policy provides specific information about protected data.
  • Do not upload, input, or analyze university data in an AI tool that is not approved by Office of Information Technology (OIT). Contracts for approved tools should ensure that the data is not utilized for training models or is isolated in a separate instance inaccessible to external parties.

Privacy

  • Be aware that most AI tools and services use input and data from users to train their models. Data entered into them, including personal data, cannot be deleted or retrieved in the future. Some tools may also access data in other open windows or applications on your device (i.e. “agentic AI”).

Procurement

  • The acquisition of generative AI tools or products that contain AI functionality, whether through purchase, subscription, or free use on behalf of AU, falls under the university's managed risk framework. University users should follow the Best Practices for Procurement of AI Tools guidance.

Do No Harm

Specific Guidance

  • Departments, units, or organizations within the university may develop their own guidance, but it must be in conformity with the general principles addressed in this document and all university policies.

Guidelines for Teaching

  • Course syllabi should clearly state policies and expectations about the use of AI, whether prohibited or permitted. Be aware that AI capabilities are increasingly integrated into other academic tools and the end user may not be fully aware. See Academic Integrity section. Faculty may consider having students affirm adherence to these policies when submitting work.
  • Instructors should recognize that the use of AI tools requires literacy about the risks and capabilities of such tools, engaging students in discussion about risks and benefits can deepen their understanding of the tools.
  • Be transparent if AI is being used in the course, including, but not limited to, generation of course materials. Instructors are accountable for the work product, decisions, and impact of the output of AI tools.
  • It is not recommended that instructors solely rely on AI to grade, assess, or give feedback on student assignments. All final decisions about grades must be human.
  • Instructors should respect and preserve students’ intellectual property rights prior to uploading any student materials to an AI tool.
  • Understand, when assigning or requiring use, that AI access amongst students can be an equity issue (paid vs. free services).
  • Instructors should also be aware of potential data usage and data privacy implications and policies [See Data Usage and Privacy sections of this document].
  • As of this date, the University does not subscribe to any AI detection tools. Additionally, the Office of Academic Integrity does not accept documentation from these tools as proof of a violation, allegations of such violations should be accompanied by other supporting evidence.

Guidelines for Research

  • Be transparent about the use of AI in your research. Clearly track and, where appropriate, disclose when, how, and which AI tools (including versions) are used, whether for data analysis, content generation, or writing assistance. Researchers are responsible for the reproducibility of their research outcomes. Remember that the standards around AI usage and disclosure are emergent and variable across disciplines. Consult discipline and publication specific citation guidance to ensure transparency meets the prevailing standard for your discipline/publication.
  • Ensure the accuracy of AI-generated content. Always verify outputs against trusted sources, especially references and factual claims. Researchers bear full responsibility for the integrity of any material included in their research.
  • Protect data privacy. Do not input sensitive or proprietary information into publicly available AI tools. Prioritize use of self-hosted AI systems or, if not available, AI systems that meet strict security standards (e.g., ISO 27001, SOC 2) and include opt-out options for data training or retention. If original data are shared for reproducibility purposes, every effort should be made to ensure that human subjects cannot be re-identified by a third-party using AI.
  • Evaluate potential bias. Critically assess ethical biases in AI-generated outputs, using automatic fairness checkers as supporting, rather than exclusive, tools. Consider how results might impact vulnerable populations or reinforce existing inequities.
  • Respect intellectual property. Attribute AI-assisted contributions appropriately and do not list AI systems as a co-author, as they cannot be held accountable for research integrity. Follow publisher and tool-specific guidelines regarding attribution, licensing, and acceptable use.
  • Do not use AI for academic tasks requiring human ethical judgment, such as peer reviews or proposal evaluations. If using AI to assist in grant writing or research compliance documentation (e.g., IRB submissions), the principal investigator must verify and disclose its involvement.
  • Use AI responsibly and ethically. Avoid using AI to generate or disseminate malicious or misleading content. Consider how data input might be reused to train future models, and use opt-out features or local models to protect sensitive information.

Guidelines for Staff

  • AI usage must be consistent with general guidelines and policies, including the Staff Personnel Policies, Data Classification Policy, or any other AU policies.
  • AI-generated outcomes must always be subject to human oversight. Decisions impacting individuals at AU - including, but not limited to, those related to employment, admissions, student affairs, financial, editorial, fact-checking, validation, sourcing/crediting, or legal - require human accountability and cannot be made solely by AI systems.
  • AI generated content should never be represented as the original work of any person in any context.
  • Consistent with the general expectation of transparency, staff should disclose use based on how the work product is being used and/or could be interpreted.
    • Staff should ensure disclosure of use is clear and accessible in all cases in which content is generated or substantially modified by AI, including text, images, audio, video, and other media.
  • All staff are bound by general guidelines for responsible use, including data usage and privacy, and are both responsible and accountable for any actions taken using AI. When in doubt, staff should not input or share data with an AI platform until they have discussed it with their supervisor or consulted with department or division leadership.

Guidelines for Students

  • Students are responsible for reviewing, understanding and adhering to the AI policies listed in course syllabi as it relates to using AI in each course. Misuse of AI or use of AI without permission or disclosure may constitute a violation of the Academic Integrity Code.
  • Students are responsible for adhering to any guidance or regulations on the use of AI tools for any extracurricular submissions or output.
  • Students are responsible for the output of any AI tool(s) they use and should be mindful of how best to use and not use AI tools. Any mistakes in the response or output generated by AI become the students’ mistakes; students should control for this by review and correction.
  • Students are partners in responsibility in achieving the learning objectives of each course and should not offload the learning to AI tools.
  • Given the inherent privacy concerns raised by the use of AI tools, students should be proactive in seeking permission before inputting the work of others into an AI tool. This also adheres to a strict reading of the AU Responsible Use of University Technology Resources Policy.
  • Students who do not have access to AI tools (paid or otherwise) should not be disadvantaged; students should address their ability to meet any course requirements directly with their instructor, department, or the Office of Dean of Students.