Effective Date: January 2026 · Last reviewed: August 2026

Purpose and Scope

This policy establishes guidelines for the use of artificial intelligence (AI) tools in all laboratory research activities, including data analysis, manuscript preparation, grant writing, and peer review. All laboratory members, as defined in the Laboratory Code of Conduct, must adhere to these standards.1,2,3

Core Principles

Laboratory AI use is guided by two principles:

  • Transparency: Full disclosure of AI applications in research outputs.
  • Accountability: Humans remain solely responsible for the accuracy and integrity of all AI-generated outputs.

Permitted Uses

AI tools may be used for the following purposes, subject to the verification and disclosure requirements below. Every permitted use below is subject to the data classification rules in Data Protection and Privacy. The activity is permitted on a public platform only where the material you paste in is Open; where it is Sensitive or Restricted, which covers anything unpublished or confidential, the same activity is permitted only in an institutionally approved secure environment. Credentials are never pasted into any AI tool, secure environment included. For writing and grant drafting assistance, a public platform may be used for general phrasing questions. Unpublished manuscripts and unsubmitted proposals are not entered into a public platform.

  1. Bioinformatics Workflows: Including sequence quality control, taxonomic classification assistance, and metagenomic binning support.5,6,7
  2. Literature Review: Summarizing literature and generating hypotheses (must be verified against primary sources).
  3. Data Visualization & Statistics: Generating code for plots or suggesting statistical approaches.
  4. Writing Assistance: Improving manuscript language, grammar, flow, and readability.
  5. Coding Assistance: Debugging scripts and generating analysis pipelines.
  6. Grant Drafting: Generating preliminary text for grant applications, adhering to sponsor guidelines.1,2,3

All uses must be documented with tool names, versions, dates, and specific applications.3,4

Prohibited Uses

The following applications are prohibited:

  1. Unattributed Content: Submitting AI-generated content as original work without substantial human contribution and disclosure.1,2,8
  2. Peer Review Violation: Using AI tools to conduct peer review of manuscripts or grant applications. Uploading confidential manuscripts or proposals to AI platforms violates confidentiality obligations and is prohibited by both NIH and NSF.9,10
  3. Data Privacy Violations: Inputting confidential, proprietary, or unpublished research data into public AI platforms (e.g., standard ChatGPT, Claude) without approved data protection measures.1,2,12
  4. Sensitive Data Exposure: Entering human subjects data, protected health information (HIPAA), student records (FERPA), or export-controlled information into unapproved AI systems.1,2,3
  5. Code Security Risks: Pasting API keys, passwords, tokens, or server credentials into any AI tool, for any purpose, including an institutionally approved secure environment.
  6. AI Authorship: Listing AI tools as authors or co-authors on publications.8,13,14,15
  7. Fabrication: Using AI to generate references, citations, or synthetic primary data (e.g., fabricating microscope images or sequencing reads).8,13,15

Data Protection and Privacy

Researchers must classify data according to UF’s Data Classification Policy before using AI tools.1,3,25

  • Public AI Platforms: (e.g., ChatGPT, Claude, Gemini) may process data classified as Open under UF’s Data Classification Policy. “De-identified” is a human subjects concept: environmental sequence data contains no personal identifiers but is still Sensitive while unpublished, and Sensitive data does not go into a public platform.2,12,23
  • Sensitive and Restricted Data: UF classifies research work in progress as Sensitive and applies Restricted to data bound by law, regulation, or contract. Both categories, which include unpublished sequences, microbiome datasets with human subjects metadata, proprietary collaborator data, and preliminary results, require an institutionally approved, secure AI environment (e.g., UF HiPerGator-RV / ResVault for regulated data), or must not be processed through AI systems at all.25,1

When in doubt, consult with the Principal Investigator before entering data into any AI tool.

Disclosure and Attribution

All AI use in research outputs must be disclosed.1,4,8,13

  • Manuscripts: Describe AI applications in the Methods section where used for data analysis or research design, and in all cases include a disclosure statement in the Acknowledgments naming the tool, version, purpose, and extent of use. ESA journals require both the in-section description and the separate Acknowledgments statement.8,13,14,15
  • Grant Applications: Follow sponsor-specific guidelines. NIH permits limited AI use but will not consider applications that are substantially developed by AI. NIH also limits the number of applications an individual principal investigator may submit per calendar year; confirm the current limit with UF Sponsored Programs before planning submissions.11 NSF prohibits its reviewers from uploading proposal content, review information, or related records to non-approved generative AI tools, and encourages proposers to indicate in the project description whether and how generative AI was used to develop the proposal.10

Authors are responsible for the accuracy, originality, and integrity of all work regardless of AI involvement.8,13,15

Quality Control and Validation

All AI-generated outputs must undergo human verification.1,4,17 Researchers must:

  1. Validate all data interpretations, statistical results, and biological conclusions.
  2. Check for algorithmic bias, including in taxonomic classifications and functional predictions.5,7
  3. Verify all citations and factual claims against the primary source; AI-generated citations may be fabricated.15
  4. Ensure reproducibility by documenting AI tool parameters, prompts, and version information.1,3,4

Training and Education

New laboratory members will receive onboarding on this policy and on discipline-specific AI guidance.3,18,19 Onboarding and the annual policy review cover current AI methods in ecology and microbiology,6,21 institutional guidance, and journal policies.20,8,13 Principal investigators and senior researchers are responsible for mentoring trainees on this policy.22

Compliance and Review

This policy aligns with UF’s AI research guidance,1,3 federal funding agency requirements,9,10,11,16 peer institutional guidance,2,12,17,23 and journal and publisher policies in ecology and microbiology.8,13,14,15 Research misconduct at UF means fabrication, falsification, or plagiarism (Policy 14-004). Most breaches of this policy are not research misconduct and are handled within the laboratory. Where conduct also involves fabrication, falsification, or plagiarism, it is referred to the UF Research Integrity Officer. On NSF awards, note that NSF 26-200 extends NSF’s own research misconduct definition to explicitly encompass AI-based tools; NSF’s definition, not UF’s, governs conduct on NSF-funded work. Confirm scope with UF RISC.1,16 This policy will be reviewed annually.

Questions and Guidance

Laboratory members with questions about this policy should consult with the Principal Investigator before proceeding. For uses involving Sensitive or Restricted data, IRB protocols, or applications not listed above, the environment must be confirmed as institutionally approved with UF’s AI working group or Office of Research before any data is entered.1,24

Policy Acknowledgment

Confirm by email, or sign a printed copy if you prefer.

Member Name (print): ___________________________________

Member Signature: ___________________________________
Date: __________________

Member Acknowledgment: All laboratory members must read and acknowledge understanding of this policy upon joining the laboratory and annually thereafter.

References

  1. University of Florida. Guidance for Researchers - AI.
  2. Northeastern University, NU-RES Compliance. Standards for the Use of Artificial Intelligence in Research at Northeastern.
  3. University of Florida. AI Technologies in Education at the University of Florida (PDF). See "Guidance for Researchers," p. 19.
  4. Helmy M, Jin L, Alhossary A, Mansour T, Pellagrina D, Selvarajoo K. Ten simple rules for optimal and careful use of generative AI in science. PLOS Computational Biology. 2025;21(10):e1013588.
  5. Zhou T, Zhao F. AI-empowered human microbiome research. Gut. 2026;75(7):1432-1446.
  6. Wang X-W, Wang T, Liu Y-Y. Artificial Intelligence for Microbiology and Microbiome Research. arXiv:2411.01098. 2024.
  7. Athanasopoulou K, Michalopoulou V-I, Scorilas A, Adamopoulos PG. Integrating Artificial Intelligence in Next-Generation Sequencing: Advances, Challenges, and Future Directions. Current Issues in Molecular Biology. 2025;47(6):470.
  8. Ecological Society of America. Artificial Intelligence (AI) Policy - Publications.
  9. NIH Office of Extramural Research. The Use of Generative Artificial Intelligence Technologies is Prohibited for the NIH Peer Review Process (NOT-OD-23-149). 2023.
  10. National Science Foundation. Notice to research community: Use of generative artificial intelligence technology in the NSF merit review process. 2023.
  11. NIH Extramural Nexus. Apply Responsibly: Policy on AI Use in NIH Research Applications and Limiting Submissions per PI (NOT-OD-25-132). 2025.
  12. Harvard University Information Technology. Generative AI Guidelines.
  13. Council of Science Editors. CSE Guidance on Machine Learning and Artificial Intelligence Tools. Science Editor. 2023;46:72.
  14. ESA Journals - Wiley. Ecology Author Guidelines.
  15. Elsevier. Generative AI policies for journals. Updated June 2026.
  16. National Institutes of Health. Research Integrity.
  17. University of Illinois Urbana-Champaign. Best Practices in Using Generative AI in Research.
  18. University of Georgia, Office of Research Integrity and Safety. Guidance on AI Use in Research. 2024.
  19. University of Washington Graduate School. Effective and Responsible Use of AI in Research. 2024.
  20. University of Florida, George A. Smathers Libraries. Artificial Intelligence: AI @ UF.
  21. Ecological Society of America. Media Tip Sheet: AI in ecology at ESA's 2026 Annual Meeting. 2026.
  22. Ecological Society of America. Code of Ethics. Amended May 2021.
  23. MIT Information Systems & Technology. Guidance for use of Generative AI tools. 2025.
  24. University of Florida. Working Group in AI Ethics and Policy.
  25. University of Florida. Data Classification Policy (Policy 12-011).