The SHARE Framework: A Practical Guide to Responsible AI Use in Research and Academia

What Is the SHARE Framework?

The SHARE Framework is a structured approach developed by York University Teaching Commons to guide responsible scholarly co-writing with generative artificial intelligence (AI). It was created to help researchers, educators, and students integrate AI into academic writing while preserving human authorship, critical thinking, and research integrity. Rather than viewing AI as an author or decision-maker, the framework positions it as a collaborative tool that supports — but never replaces — the expertise and accountability of human researchers.

Within the context of AI in research and AI in academia, the SHARE Framework encourages transparent, ethical, and reflective use of generative AI throughout the research and writing process. It emphasizes that researchers remain responsible for the accuracy of facts, the quality of analysis, and the originality of scholarly contributions, even when AI is used to assist with drafting, editing, or organizing ideas. The framework therefore promotes responsible AI, ensuring that AI-assisted research enhances productivity without compromising academic integrity. As generative AI becomes increasingly common in higher education, the SHARE Framework offers practical guidance for balancing technological innovation with ethical scholarly practice and accountable human authorship.

Why the SHARE Framework Matters in Research and Academia

Generative AI is rapidly transforming research and higher education by helping scholars generate ideas, summarize literature, improve writing quality, and streamline academic workflows. These capabilities increase productivity and reduce the time required for routine writing tasks. However, AI-generated content also introduces significant risks, including fabricated citations, factual inaccuracies, biased outputs, and the potential erosion of independent critical thinking. Without appropriate oversight, researchers may unknowingly incorporate misleading or unverifiable information into scholarly work.

The SHARE Framework addresses these challenges by promoting a responsible model of AI-assisted research that keeps human judgment at the center of the scholarly process. Instead of allowing AI to determine arguments or conclusions, researchers are encouraged to critically evaluate AI-generated content, verify evidence against reliable scholarly sources, and maintain full responsibility for published work. This approach strengthens academic integrity while allowing researchers to benefit from advances in generative AI. For example, York University Teaching Commons recommends the SHARE Framework as guidance for scholarly co-writing, emphasizing accountable authorship, transparency, and ethical AI use throughout the writing process. By combining technological efficiency with rigorous scholarly standards, the framework supports responsible innovation across universities and research institutions.

The Five Principles of the SHARE Framework

To use the SHARE Framework for scholarly co-writing effectively, researchers should treat generative AI as a supportive partner within a clearly human-led process. The following five principles provide practical guidance for responsible AI-assisted research, academic writing with AI, and ethical decision-making throughout the research lifecycle.

1. Support

The Support principle positions generative AI as a research and writing assistant rather than an author. Researchers can use AI to brainstorm research questions, create outlines, summarize literature, improve grammar, and refine language for clarity and readability. These tasks can make academic workflows more efficient while allowing scholars to focus on analysis, interpretation, and original contribution. However, researchers must retain ownership of all ideas, arguments, methods, and conclusions. AI-generated suggestions should be treated as draft material that requires human review, revision, and verification before inclusion in scholarly work. The principle therefore encourages responsible AI in higher education by emphasizing that AI should enhance — not replace — critical thinking and disciplinary expertise.

Example: A postgraduate researcher uses AI to summarize several peer-reviewed articles, then independently checks the findings, compares the sources, and writes the final literature review using their own synthesis and interpretation.

2. Human-Centered Collaboration

The Human-Centered Collaboration principle emphasizes that researchers and AI work together, but humans remain in control of the scholarly process. Effective collaboration begins with purposeful prompting, followed by careful evaluation, editing, and contextualization of AI-generated content. Human expertise is essential for selecting relevant evidence, applying disciplinary standards, interpreting results, and ensuring that arguments are appropriate for the intended academic audience. Researchers should therefore view AI outputs as starting points for discussion and refinement rather than finished scholarly products. This principle supports ethical AI-assisted scholarly writing by preserving human judgment, creativity, and accountability at every stage of research and publication.

Example: A research team uses AI to draft background text for a grant proposal, while subject-matter experts revise the literature review, design the methodology, verify references, and finalize the analysis before submission, ensuring the proposal reflects their professional expertise and scholarly standards.

Source: York University Teaching Commons — SHARE: A Framework for Scholarly Co-Writing with Generative AI.

3. Accountability

The Accountability principle states that researchers remain fully responsible for every aspect of their published work, regardless of whether AI was used during drafting or editing. Authors must verify factual claims, check citations against original sources, confirm that interpretations are accurate, and ensure that AI-generated text does not introduce errors or misleading statements. Transparency is also important: researchers should disclose AI use whenever required by their institution, funder, conference, or journal. By maintaining clear records of how AI was used, scholars can demonstrate compliance with academic and publishing standards while protecting the integrity of the research process. This principle reinforces responsible AI use in research by linking AI assistance to human accountability rather than machine authority.

Example: Before submitting a manuscript, authors independently verify all references and confirm that they — not the AI system — accept responsibility for the accuracy, originality, and integrity of the final paper.

4. Reflection

The Reflection principle encourages researchers to critically evaluate AI-generated outputs before relying on them in academic work. Generative AI can produce confident but inaccurate statements, omit important context, or reflect biases present in training data. Researchers should therefore compare AI-generated summaries, explanations, and recommendations with peer-reviewed literature, authoritative databases, and primary sources. Reflection also involves assessing whether AI is genuinely improving the quality of the research and refining prompts when outputs are incomplete or misleading. By pausing to question, verify, and revise AI-generated material, scholars strengthen both the reliability of their findings and their own critical-thinking skills. This principle is central to AI research ethics because it treats verification as an ongoing scholarly responsibility rather than a final proofreading step.

Example: A doctoral student checks an AI-generated literature summary against articles retrieved from Scopus and Web of Science before incorporating any information into the dissertation.

5. Ethics

The Ethics principle requires researchers to use generative AI in ways that uphold academic integrity, protect sensitive information, and respect intellectual property. Scholars should avoid plagiarism, properly acknowledge AI assistance when required, and ensure that confidential, unpublished, or personally identifiable research data are not entered into systems that could compromise privacy or security. Researchers must also follow university policies, funder requirements, and publisher guidelines governing AI use in scholarly communication. Ethical practice extends beyond compliance: it involves considering fairness, transparency, and the broader impact of AI on teaching, research, and public trust. By embedding ethical judgment throughout the research process, the SHARE Framework supports responsible AI in academia while preserving the credibility of scholarly work.

Example: A researcher follows institutional AI guidance, excludes confidential participant data from AI tools, and discloses permitted AI assistance in the manuscript’s acknowledgments or methods section.

Benefits of the SHARE Framework

The SHARE Framework helps researchers integrate generative AI into academic work while maintaining high standards of scholarship and research integrity. Its structured approach offers several advantages for researchers, educators, and higher education institutions.

Key benefits include:

• Promotes responsible AI adoption by encouraging researchers to use AI as a support tool rather than a substitute for human expertise.
• Strengthens academic integrity through verification, transparency, and ethical decision-making.
• Improves research quality by encouraging critical evaluation of AI-generated content before publication.
• Enhances transparency by supporting appropriate disclosure of AI use in scholarly writing.
• Supports ethical scholarly co-writing through clear human oversight and accountability.
• Builds trust among publishers, institutions, and readers by reinforcing author responsibility.
• Encourages productive human-AI collaboration while preserving originality and critical thinking.

Practical example: York University encourages responsible scholarly co-writing practices that help postgraduate students use AI to improve writing efficiency without compromising academic standards.

Challenges of Applying the SHARE Framework

Although the SHARE Framework provides practical guidance, implementing it consistently can present several challenges.

Common challenges include:

• Limited AI literacy among researchers and students.
• Overreliance on AI-generated content.
• Maintaining originality while using AI-assisted writing tools.
• Differences in university and publisher AI policies.
• Additional time required to verify AI-generated facts and references.
• Rapid advances in generative AI that continually change best practices.

Practical example: A researcher may save time drafting a literature review with AI but spend additional hours verifying references because generative AI can produce inaccurate or fabricated citations, making human verification essential before publication.

Best Practices for Applying the SHARE Framework

Researchers can maximize the value of the SHARE Framework by following these best practices:

• Verify every AI-generated fact, quotation, and reference.
• Cross-check citations using trusted scholarly databases such as Scopus, Web of Science, or Google Scholar.
• Maintain independent analysis and critical thinking throughout the research process.
• Disclose AI use whenever required by publishers or institutions.
• Follow university and journal AI policies.
• Protect confidential, unpublished, and sensitive research data.
• Continuously improve AI literacy as generative AI technologies evolve.

Case Study: Applying the SHARE Framework in Scholarly Writing

A postgraduate researcher preparing a journal manuscript uses generative AI to brainstorm research questions, organize an outline, and improve sentence clarity. Before incorporating any AI-generated content, the researcher verifies every citation against peer-reviewed literature, rewrites the text using independent analysis, and removes unsupported claims. Confidential research data are never entered into the AI system, and AI assistance is disclosed according to the target journal’s submission guidelines. By combining Support, Human-Centered Collaboration, Accountability, Reflection, and Ethics, the researcher produces a transparent, high-quality manuscript that meets academic and publishing standards.

Conclusion

The SHARE Framework provides researchers with a practical foundation for using generative AI responsibly throughout the scholarly writing process. By emphasizing Support, Human-Centered Collaboration, Accountability, Reflection, and Ethics, it reinforces that AI should enhance — not replace — human expertise. As AI becomes increasingly integrated into research and higher education, frameworks such as SHARE promote transparency, academic integrity, and responsible innovation, enabling researchers to benefit from AI-assisted scholarship while preserving the credibility, originality, and trustworthiness of academic research.