Teaching & Learning
- For graduate students, AI should not be used to complete graded assignments
- The guidance says academic work must reflect the student's own thinking and integrity, and substantial AI contribution must be acknowledged or cited
Your academic work should always reflect your own
thinking and integrity.
✓ Always review, evaluate, and verify AI-generated outputs.
✓ If AI substantially contributes, acknowledge or cite it:
Example: “Assisted by ChatGPT, OpenAI (2025)”
Avoid using AI to:
➢ Complete graded assignments
➢ Generate text without attribution
- The guidance also requires users to review, evaluate, and verify AI outputs and to acknowledge or cite substantial AI contribution
- For graduate students, AI tools are presented as allowable support for academic study and research support tasks such as brainstorming, language editing, summarizing articles, generating citations, and generating code
🤖 AI tools can support research in:
➢ Generating ideas – Brainstorming– Language editing – Summarizing articles – Generating citations or
code
✓ Always review, evaluate, and verify AI-generated outputs.
✓ If AI substantially contributes, acknowledge or cite it:
Example: “Assisted by ChatGPT, OpenAI (2025)”
- The same graduate guidance requires review and verification of AI outputs and says substantial AI contribution should be acknowledged or cited
- For graduate students, AI tools are described as able to support research by generating code, but the university does not provide a specific university-wide rule for programming assignments
🤖 AI tools can support research in:
➢ Generating ideas – Brainstorming– Language editing – Summarizing articles – Generating citations or
code
✓ Always review, evaluate, and verify AI-generated outputs.
✓ If AI substantially contributes, acknowledge or cite it:
Example: “Assisted by ChatGPT, OpenAI (2025)”
Research
- In research, AI may be used across writing and manuscript stages, including literature review, drafting prose, formatting references, and preparing lay summaries
- Researchers remain fully responsible for the work, human judgment must not be replaced, and AI assistance in drafting or publication must be acknowledged, cited, or disclosed according to journal requirements
Researchers are fully responsible and accountable for the use of AI in research, ensuring that AI
tools do not replace human judgment. Human oversight must guide all critical decisions,
interpretations, and ethical considerations.
GenAI may be used through various stages of the research process from ideation, formulating a
research hypothesis or question, writing a literature review, data gathering and analysis to
writing and reviewing manuscript.
- Refine language for funding narratives, writing lay summaries for community partners.
- State in acknowledgements that GenAI assisted drafting.
- GenAI helps turn structured notes into fluent prose, formats references in APA, and drafts
alt-text for figures.
- In case of publication, take into consideration the requirements of the
publisher/journal on disclosure or citation of GenAI tools (check Appendix for examples).
- Cite the tool as software, for example, “Portions of the introduction were refined using
ChatGPT, OpenAI GPT-4o, May 2025 release”, in case the journal favors citation to
disclosure.
- Disclose the GenAI, instead of citing it, in the cases where the journals favor disclosure to
citation.
- The university allows AI use in research data collection, generation, management, and analysis, but only with substantial safeguards
- Synthetic data must be kept separate and documented, sensitive research data must stay on secure AUB servers or approved encrypted/private systems, anonymization and human review are required before upload, and AI-generated statistics must be rerun with trusted software
Researchers must prioritize the privacy and confidentiality of participant data when using AI tools
in research. Confidential data must not be shared with AI systems unless supported by
appropriate data protection agreements.
- Create synthetic data sets to test your data collection/extraction tools.
- Use GenAI as a chatbot interviewer.
- Keep synthetic data separate from observational data: tag synthetic records clearly so
they do not merge with real data.
- Don’t use AI for social platforms scraping without consent from subjects.
- Document clearly how the data was generated.
- Use GenAI only inside secure AUB servers or approved encrypted clouds. Encryption at
rest and in transit is mandatory. Research involving sensitive data (patient, financial,
political) must use secure in-house or approved private AI systems, not public GenAI APIs.
- Ensure that data is anonymized by a human data steward, before uploading it into the
approved AI tool.
- Ensure human review of anonymized data to prevent data leakage.
- GenAI can draft code snippets, regular expressions, or commentary on statistical output.
- Re-run GenAI generated statistics with trusted software.
- Disclose model detail needed for replication, in the methods section: model names,
parameters, and system prompts.
- AUB requires researchers to consider AI ethics throughout the research process and to follow institutional, national, and funder rules
- For ethics and regulatory submissions, AI-generated drafts may be used for documents like consent forms, but prompts and outputs should be attached to IRB materials for traceability, personal data must not be entered into public models without consent, and proposal AI use must be acknowledged while full proposals and intellectual property content are barred from AI systems that retain or train on uploaded data
Researchers must consider the ethical implications of AI use at every phase of their work. This
includes avoiding bias in AI-generated content, ensuring fairness and inclusivity, protecting
vulnerable populations, and maintaining the integrity of the research process. The use of AI
should uphold the core ethical principles of respect, beneficence, and justice.
Researchers must adhere to institutional, national, and funder policies on the responsible use of
AI, especially concerning intellectual property (IP) rights, data protection, and research ethics.
Researchers should review any policies and requirements that discuss the usage of GenAI tools
in their research. Such policies will indicate whether the use of AI is permitted and how you are
required to disclose the usage of such tools.
- Generate first drafts of participant information sheets, consent forms, or data-sharing
agreements, then review wording against GDPR lawful-basis articles and HIPAA
protections for human subject research.
- Attach the full GenAI prompt and output, including API calls, to your IRB packet; many
committees now request this for traceability.
- Don’t feed personal data into a public model without consent from the subject.
- When applicable, the use of AI in the preparation of proposals must be clearly acknowledged.
- It is strictly prohibited to upload full proposals or intellectual property content to any AI
platform that stores, retains, or uses uploaded data to train its models or databases. This
includes using such platforms for editing, reviewing, or identifying potential reviewers. Only
AI tools that function as local or secure editors—without retaining or transmitting data
externally—may be used, provided institutional data protection standards are met.
Academic Integrity
- AUB requires transparency about AI use in research and graduate academic work
- Researchers must document when, how, and which tools were used, and may need to record prompts, methods details, acknowledgements, or journal-specific citation/disclosure statements; graduate students are told to acknowledge or cite substantial AI contribution
Researchers must document when, how and which AI tools are used, along with their influence
on the research and dissemination process, to ensure compliance with academic and research
integrity standards and to ensure reproducibility. In line with open research practices, it is good
practice to record the prompts used to generate AI outputs.
- Record each session, including prompts, date, tool name, and version.
- Document runs with the exact model version, decoding parameters, full prompts/context
(including any retrieved data), and final outputs.
- State in acknowledgements that GenAI assisted drafting.
- Disclose model detail needed for replication, in the methods section: model names,
parameters, and system prompts.
- In case of publication, take into consideration the requirements of the
publisher/journal on disclosure or citation of GenAI tools (check Appendix for examples).
- Cite the tool as software, for example, “Portions of the introduction were refined using
ChatGPT, OpenAI GPT-4o, May 2025 release”, in case the journal favors citation to
disclosure.
- Disclose the GenAI, instead of citing it, in the cases where the journals favor disclosure to
citation.
- Deposit prompts and raw outputs as supplementary material if requested by journal.
✓ If AI substantially contributes, acknowledge or cite it:
Example: “Assisted by ChatGPT, OpenAI (2025)”
- When applicable, the use of AI in the preparation of proposals must be clearly acknowledged.
- However, the sources do state prohibited uses of AI in graduate academic work and in research proposal handling
- The provided sources do not define the university's use of AI detection tools or formal misconduct penalties specifically tied to AI detection
Avoid using AI to:
➢ Complete graded assignments
➢ Generate text without attribution
➢ Generate or falsify data
- It is strictly prohibited to upload full proposals or intellectual property content to any AI
platform that stores, retains, or uses uploaded data to train its models or databases.
- Reviewers are strictly prohibited from uploading proposals or any related materials to
generative AI platforms, particularly those that store, process, or use uploaded data for model
training or external purposes, as this may violate the authors’ confidentiality and proprietary
rights and, where the paper contains personally identifiable information, may breach data
privacy rights.
Institutional & Administrative
- For research administration, reviewers are prohibited from uploading proposals to generative AI platforms that store or use data externally, and limited non-confidential information may be used to identify reviewers
- The sources define AI-related responsibilities for researchers, principal investigators, custodians, and proposal reviewers, but do not set a general university-wide policy for faculty teaching or staff administrative uses such as grading, lesson planning, or recommendation letters
Researchers are fully responsible and accountable for the use of AI in research, ensuring that AI
tools do not replace human judgment. Human oversight must guide all critical decisions,
interpretations, and ethical considerations.
- To identify potential reviewers using AI tools, custodians may only upload limited, non-confidential information such as proposal titles, keywords, and references.
- Reviewers are strictly prohibited from uploading proposals or any related materials to
generative AI platforms, particularly those that store, process, or use uploaded data for model
training or external purposes, as this may violate the authors’ confidentiality and proprietary
rights and, where the paper contains personally identifiable information, may breach data
privacy rights.
- AUB sets explicit data protection rules for AI use in research and proposal preparation
- Users should assume web-based AI inputs may become training data, confidential data must not be shared without proper agreements, sensitive research data must be confined to secure AUB servers or approved encrypted/private systems, and only local or secure editors that do not retain or transmit proposal content externally may be used
While some GenAI services offer a “no training use” policy as part of enterprise
agreements, by default, researchers should first assume that any data or information entered
into web-based GenAI could become part of its training data and can thus be intentionally or
unintentionally (post a leak or an incident) recovered by third parties.
Researchers must prioritize the privacy and confidentiality of participant data when using AI tools
in research. Confidential data must not be shared with AI systems unless supported by
appropriate data protection agreements.
- Use GenAI only inside secure AUB servers or approved encrypted clouds. Encryption at
rest and in transit is mandatory. Research involving sensitive data (patient, financial,
political) must use secure in-house or approved private AI systems, not public GenAI APIs.
- Ensure that data is anonymized by a human data steward, before uploading it into the
approved AI tool.
- It is strictly prohibited to upload full proposals or intellectual property content to any AI
platform that stores, retains, or uses uploaded data to train its models or databases. This
includes using such platforms for editing, reviewing, or identifying potential reviewers. Only
AI tools that function as local or secure editors—without retaining or transmitting data
externally—may be used, provided institutional data protection standards are met.
- The documents emphasize research integrity, responsible use, and future integration of lessons learned about AI's impact on teaching, learning, and academic research
- AUB has issued initial institutional guidance for responsible AI use in research and states that AI guidance will continue to evolve through institutional monitoring and community feedback
In line with our commitment to upholding research integrity amid the emergence of Generative
AI (GenAI), this document offers initial guidelines on the responsible usage of GenAI in research.
Raising awareness on the opportunities, limitations, and responsible use of GenAI is important
and is expected to become an integral part of the future work of our institution. Since GenAI is
evolving rapidly, AUB will keep on monitoring the technological developments and integrate
feedback from the AUB community to update the guidelines as we learn more about the impact
of GenAI on teaching, learning, and academic research.
DocuMark: Responsible AI Use for Academic Integrity
Knowing your institution's AI policy is step one. DocuMark helps enforce it fairly by empowering universities to manage AI-generated content, prevent cheating, and support student writing through responsible AI use.
Common Questions About American University of Beirut's AI Policies
American University of Beirut has defined AI policies in 11 of 12 categories, with an overall coverage score of 92%.
AUB requires transparency about AI use in research and graduate academic work. Researchers must document when, how, and which tools were used, and may need to record prompts, methods details, acknowledgements, or journal-specific citation/disclosure statements; graduate students are told to acknowledge or cite substantial AI contribution.
The provided sources do not define the university's use of AI detection tools or formal misconduct penalties specifically tied to AI detection. However, the sources do state prohibited uses of AI in graduate academic work and in research proposal handling.
AUB sets explicit data protection rules for AI use in research and proposal preparation. Users should assume web-based AI inputs may become training data, confidential data must not be shared without proper agreements, sensitive research data must be confined to secure AUB servers or approved encrypted/private systems, and only local or secure editors that do not retain or transmit proposal content externally may be used.