Search across guide titles, topics, keywords, and content throughout the Researcher's Field Guide.
Showing 251–275 of 872 results
An AI tool does not need to reveal every technical detail to be useful for research. It becomes too opaque when missing information prevents researchers from adequately evaluating, verifying, governing, or defending the role it will play.
Read Guide →AI use in research is rarely governed by a single policy. Researchers may need to comply simultaneously with laws, university rules, ethics approvals, funder conditions, journal policies, and research-team agreements.
Read Guide →Research funders can set their own requirements for AI use in funding applications, assessment, funded projects, and reporting. These rules may be stricter than your university’s general AI policy.
Read Guide →A research ethics committee may restrict or require changes to AI use when it affects participants, consent, privacy, confidentiality, risk, or an approved research protocol.
Read Guide →Your university may permit an AI tool while a journal restricts how it can be used in submitted work. Institutional permission and publication requirements govern different parts of the research process.
Read Guide →University and journal AI policies can impose different requirements, but not every difference is a true conflict. Learn how to identify which rules apply and what to do when both cannot be satisfied as written.
Read Guide →If a journal has no clear AI policy, do not treat silence as permission or prohibition. Check publisher-level rules, existing editorial requirements, and the nature of your AI use before deciding whether clarification is needed.
Read Guide →AI policies may change while research is underway. Researchers should identify exactly what changed, determine when and to whom the new rule applies, adjust future activity where necessary, and document how earlier AI use was handled.
Read Guide →When researchers from several institutions collaborate, there may be no single AI policy covering the entire project. Researchers need to map which institutional, ethical, contractual, funder, and data rules apply to each activity.
Read Guide →Long research projects can outlast the AI policies that existed when they began. A simple policy register, scheduled reviews, and event-triggered checks can keep the project current without constant monitoring.
Read Guide →Entering research information into an AI tool is not automatically safe or unsafe. The risk depends on what you submit, whose information it contains, the tool's data practices, and the rules governing your research.
Read Guide →Unpublished research is not automatically confidential, but entering it into a public AI tool can create disclosure, privacy, ownership, and policy concerns. Check what the material contains and whether you are authorized to submit it before doing so.
Read Guide →Do not assume that a confidential research proposal can be uploaded to an AI tool simply because you wrote it. Proposal content may be governed by confidentiality, institutional, funder, collaborator, intellectual-property, and data-handling requirements.
Read Guide →An unpublished manuscript is not automatically prohibited from AI use, but uploading the complete document can expose far more information than a writing task requires. Check confidentiality, co-author rights, research data, publisher rules, and the AI service before submitting it.
Read Guide →Peer-review manuscripts and associated materials are confidential information entrusted to reviewers and editors for a limited purpose. Do not upload them to an external AI tool unless the journal explicitly permits the use and the required confidentiality can be assured.
Read Guide →A confidential grant application entrusted to you for review is not ordinary research material. Some major funders expressly prohibit reviewers from uploading applications, proposal content, critiques, or related review information to unapproved generative AI systems.
Read Guide →Having access to a collaborator's unpublished work does not automatically give you authority to submit it to an AI provider. Permission, confidentiality, ownership, research agreements, and the nature of the information all matter.
Read Guide →The safest default is simple: do not put confidential, sensitive, restricted, or third-party unpublished research into an unapproved AI tool. Use it only after you can establish authorization, appropriate safeguards, and a genuine need to disclose the information.
Read Guide →Uploading research participant data to an AI tool is not automatically permissible simply because the tool can process it. Researchers should first establish whether the proposed processing is authorized, necessary, appropriately protected, and consistent with consent, ethics approval, institutional policy, and applicable privacy law.
Read Guide →De-identification can reduce privacy risk, but ethical research involves more than preventing someone from attaching a name to a record. AI use may still raise questions about consent, secondary use, group harms, fairness, confidentiality, and whether participants could reasonably have expected their data to be used this way.
Read Guide →Removing names does not necessarily prevent research participants from being identified. AI and other computational methods can exploit combinations of attributes, link records with external information, or infer sensitive characteristics that were not obvious in the original dataset.
Read Guide →Research interview transcripts can contain far more identifying information than participant names. Before uploading transcripts to generative AI, researchers should assess consent, ethics approval, contextual identifiers, provider data practices, and whether the service is authorized for the data.
Read Guide →Survey responses may look less sensitive than interviews, but structured variables, free-text answers, metadata, and combinations of demographics can still identify participants. AI use should depend on the actual data and the conditions governing their processing.
Read Guide →Clinical research data may contain health information, direct identifiers, rare conditions, dates, images, genomic information, and other highly sensitive records. AI processing may be possible in approved environments, but researchers should verify the specific legal, ethical, institutional, contractual, and technical requirements before transmitting the data.
Read Guide →Research images may identify participants through faces, distinctive physical features, surroundings, metadata, or information visible within the image. Researchers should establish whether AI processing is authorized and whether the image can be minimized or transformed before uploading it.
Read Guide →