91´ÎÔª

Data Insecurity (Privacy Violations)

Using AI tools—especially cloud-based or public-facing platforms—can inadvertently expose sensitive, proprietary, or personally identifiable information. Uploading confidential data without proper safeguards may violate university policy and federal regulations. In some cases, such misuse may rise to the level of research misconduct.Ìý

Falsehoods and Inaccuracies in AI-Generated Content

AI systems may produce text or analyses that appear authoritative but are factually incorrect or misleading. Without rigorous verification, these errors can compromise the validity of research results and undermine scientific integrity.Ìý

Fabricated or Missing Citations

Generative AI tools often invent sources or fail to attribute ideas correctly. Improper attribution risks ethical violations, plagiarism allegations, reputational damage, and legal consequences.Ìý

Data Provenance and Documentation Gaps

When AI is used to analyze or generate data, it can obscure the provenance of results. Researchers must maintain transparent documentation of inputs, data transformations, model parameters, and how AI tools were used in the research process.Ìý

Noncompliance with Funder or Publisher Guidelines

Policies on AI use vary widely. Most funders prohibit AI use in peer review. Some allow AI-assisted drafting with disclosure. Researchers must review agency-specific guidance carefully and document AI use where required.Ìý

Research Integrity and Norm Evolution

The norms and ethics surrounding AI are still developing. Researchers have a professional responsibility to stay informed about emerging standards, regulatory guidance, and evolving best practices in their field.Ìý

Writing Support for Proposals

AI tools can improve writing clarity, consistency, and productivity. Potential applications include:Ìý

• Editing draft narratives (especially when peer feedback is limited)Ìý
• Generating project summaries or lay abstractsÌý
• Outlining proposals or structuring sections prior to beginning writingÌý
• Analyzing tone and coherence in complex argumentsÌý

Telling Your Research Story

Communicating complex research in accessible language is critical. AI can help clarify technical content, support public outreach, and tailor messaging for specific audiences (e.g., social media, press releases, stakeholder reports).Ìý

Preliminary Analysis of Large Data Sets

With proper validation, AI can help identify early patterns in large or unstructured datasets, supporting hypothesis generation and initial exploration.Ìý

Early-Stage Literature Review

AI can support the initial phases of a literature review by identifying key concepts, summarizing known work, and surfacing potential gaps. However, it must be supplemented by traditional methods to verify citations and assess quality.Ìý

Identifying Gaps in Literature or Research Directions

Some tools can help map research landscapes and suggest novel angles—useful for framing new grant proposals or exploratory studies.Ìý

Proposal Development with AI

• If allowed by the funder, researchers may choose to use AI tools (e.g., language models) to draft or refine parts of grant proposals. If doing so: Review and edit all AI-generated text for accuracy, tone, and alignment with the funder’s expectations.Ìý
• Be transparent if the funder requires disclosure.Ìý
• Avoid using AI to generate data, citations, or any content that you cannot fully validate or attribute.Ìý