This page is part of MIT's Guidelines for Ethical Authorship and Publication Practices.
AI tools are increasingly used across research and scholarship at MIT to support creativity, efficiency, and discovery. In this context, AI includes tools such as generative AI, machine learning models, and automated analysis systems that assist with tasks like data analysis, coding, writing, visualization, or idea development. As MIT continues to develop more detailed guidance in this rapidly evolving area, the information below offers practical interim guidance for the community.
Using AI Responsibly: Do’s and Don’ts
When used thoughtfully and transparently, AI can be a valuable support tool. AI should support—not replace—scholarly judgment, originality, and accountability. Common and appropriate uses include:
- Brainstorming ideas, outlines, or research questions
- Assisting with coding, data analysis, visualization, or debugging
- Editing for clarity, grammar, or style
- Summarizing literature or organizing notes
- Supporting routine or repetitive research tasks
AI should not be used in ways that undermine research integrity or violate MIT policies, sponsor or peer review requirements, or journal expectations. Examples of inappropriate use include:
- Presenting AI-generated content, data, or analysis as one’s own without disclosure
- Using AI to fabricate, falsify, or manipulate data, results, or citations
- Uploading confidential, proprietary, sensitive, or controlled research data into AI tools that are not approved for such use
- Using AI in ways that violate sponsor or peer reviewer requirements, journal policies, or course or program expectations
Disclosure Matters
Transparency is essential. If AI tools are used in a meaningful way—for example, in writing, analysis, image generation, or code development—this use should be disclosed as appropriate to the context, such as in publications, theses, or discussions with advisors, collaborators, or instructors.
Choosing and Vetting AI Tools at MIT
Not all AI tools are appropriate for research use. Before using third-party AI platforms, researchers should consider data privacy, security, and compliance requirements, and avoid entering sensitive or restricted information into unapproved tools. MIT’s Information Systems & Technology (IS&T) office provides guidance on evaluating AI tools for research use and can help assess security, privacy, and policy considerations. Researchers are encouraged to consult IS&T before adopting new tools and may request assistance by contacting ai-guidance@mit.edu.
A Shared Responsibility
Members of the MIT community remain fully responsible for the accuracy, integrity, and originality of their work—even when AI tools are used. AI systems can produce errors or “hallucinations,” so all AI-assisted outputs must be carefully reviewed and verified. Failure to do so could result in an allegation that you reported falsified or fabricated information. Responsible AI use requires good judgment, transparency, and awareness of limitations. By using AI thoughtfully, the MIT community can benefit from these tools while upholding high standards of research integrity and trust.