Artificial Intelligence in Research
Participant recordings can reveal identity through voices, faces, speech, surroundings, and metadata. Before using AI for transcription or analysis, researchers should determine whether the recording itself may be disclosed to the service, not merely whether the resulting transcript will be de-identified.
Read Guide →
Artificial Intelligence in Research
Informed consent does not necessarily need to name every software tool used in research, but AI processing may warrant disclosure when it materially changes how participant data are handled, who receives them, what is inferred, or the risks participants face.
Read Guide →
Artificial Intelligence in Research
Existing participant consent may sometimes be broad enough to accommodate a later AI-assisted method, but researchers should not assume that generic permission for “analysis” or “future research” authorizes materially different processing. The actual consent, protocol, data, purpose, risks, and applicable governance requirements need to be assessed.
Read Guide →
Artificial Intelligence in Research
Introducing AI into an approved study may be a protocol change when it alters participant-data processing, confidentiality, external recipients, analysis, consent, or risk. Researchers should let the responsible ethics authority determine whether formal amendment is required rather than deciding that AI is merely another software tool.
Read Guide →
Artificial Intelligence in Research
If participant data were already processed with an unauthorized AI tool, stop further processing, preserve the facts, notify the appropriate institutional channels, assess what data were exposed and what the provider did with them, and follow applicable incident and breach procedures. Quietly deleting the chat is not an adequate response.
Read Guide →
Artificial Intelligence in Research
AI should not be used with participant data merely because it makes a research task faster. When the same purpose can be achieved with less sensitive data, an approved conventional method, or a safer processing environment, the additional privacy exposure may be difficult to justify.
Read Guide →
Artificial Intelligence in Research
Generative AI can be useful for producing, expanding, and recombining possible research ideas. Its strongest role is as an ideation partner, not as an authority on whether an idea is novel, important, feasible, or supported by the literature.
Read Guide →
Artificial Intelligence in Research
Generative AI can help researchers notice possible problems, tensions, contradictions, and unexplained phenomena worth investigating. It cannot establish from plausible language alone that a research problem genuinely exists or deserves study.
Read Guide →
Artificial Intelligence in Research
Generative AI can make research-topic brainstorming faster and more varied, especially when you give it meaningful constraints. The strongest approach is not to ask for a long list of titles, but to use AI to explore different directions before selecting and validating a topic yourself.
Read Guide →
Artificial Intelligence in Research
Generative AI can help researchers notice possible gaps, inconsistencies, neglected perspectives, and unanswered questions. It cannot establish a genuine research gap unless the relevant literature has been searched and evaluated adequately.
Read Guide →
Artificial Intelligence in Research
When AI says a topic is understudied or a gap exists, it is making a claim about the literature. That claim becomes evidence only after researchers verify it against an adequate and appropriately searched body of scholarship.
Read Guide →
Artificial Intelligence in Research
There is no single universal rule saying that whoever receives an AI-generated research idea automatically “owns” it. Copyright, patent law, contracts, institutional rules, confidentiality, and scholarly credit address different aspects of ownership and may lead to different answers.
Read Guide →
Artificial Intelligence in Research
An AI-assisted research idea becomes meaningfully the researcher’s own through substantive intellectual work: questioning it, grounding it in literature, reshaping it, making consequential decisions, and taking responsibility for the resulting research.
Read Guide →
Artificial Intelligence in Research
Generative AI can help you draft, compare, and revise possible research questions, but a polished question is not necessarily a defensible one. Learn how to use AI without outsourcing the substantive decisions that define your study.
Read Guide →
Artificial Intelligence in Research
An overly broad research question can make a study difficult to design and complete. Learn how generative AI can help identify unnecessary breadth, compare narrower formulations, and preserve the original research purpose.
Read Guide →
Artificial Intelligence in Research
A research topic identifies what interests you, but a researchable question establishes what you intend to investigate. Learn how generative AI can help bridge that gap without inventing a research problem or assuming an unverified knowledge gap.
Read Guide →
Artificial Intelligence in Research
Generative AI can suggest research questions, variables, hypotheses, and methods, but it cannot assume responsibility for the decisions that define a study. Learn which intellectual judgments must remain with the researcher and how to use AI without surrendering control of the inquiry.
Read Guide →
Artificial Intelligence in Research
Generative AI can help researchers discover scientific papers, clarify search concepts, and explore unfamiliar research areas. Its usefulness depends on how it retrieves evidence, how results are verified, and whether the search process can be documented.
Read Guide →
Artificial Intelligence in Research
Generative AI-powered discovery tools may identify relevant studies that traditional keyword searches overlook by recognizing conceptual relationships rather than relying solely on matching words. However, semantic similarity introduces its own limitations, making verification and complementary database searching essential.
Read Guide →
Artificial Intelligence in Research
Generative AI can suggest synonyms, alternative terminology, spelling variants, and related research concepts to improve literature search strategies. The key is distinguishing genuine synonyms from broader concepts and validating suggestions against actual scholarly records.
Read Guide →
Artificial Intelligence in Research
Generative AI can help researchers convert research questions into Boolean search strings by organizing concepts, synonyms, and search operators. However, AI-generated queries must be checked for conceptual accuracy, database compatibility, and retrieval performance before they are used in a literature review.
Read Guide →
Artificial Intelligence in Research
Generative AI can suggest subject headings and controlled vocabulary terms for scholarly database searches, but plausible terminology is not necessarily an authorized indexing term. Researchers must verify headings in official thesauri and evaluate their meaning, hierarchy, and retrieval behavior.
Read Guide →
Artificial Intelligence in Research
Generative AI can suggest potentially seminal research papers, but familiarity, citation counts, and confident recommendations do not establish a study's historical or intellectual significance. Researchers should verify the publications and examine their actual contributions to the development of the field.
Read Guide →
Artificial Intelligence in Research
Generative AI can suggest influential research papers, but identifying the single most important publication in a field requires defining what importance means. Citation counts, originality, theoretical influence, and practical impact may point to different studies.
Read Guide →
Artificial Intelligence in Research
Generative AI can help identify recently published scientific research, particularly when connected to current scholarly retrieval sources. However, finding the most recent studies requires distinguishing publication dates from indexing dates, verifying source coverage, and accounting for delays in database updates.
Read Guide →
