Artificial Intelligence in Research
AI-assisted literature synthesis becomes unsuitable when researchers cannot reliably verify its claims, source attributions, or interpretations. Learn which errors are manageable, which undermine scholarly conclusions, and when to abandon an AI-generated synthesis rather than repeatedly correcting it.
Read Guide →
Artificial Intelligence in Research
Generative AI can assist with several stages of a systematic review, but its outputs require validation and appropriate human oversight. Learn where AI can be useful, what risks researchers must manage, and how to use it without compromising review integrity.
Read Guide →
Artificial Intelligence in Research
Generative AI can support systematic review tasks ranging from search preparation and screening to data extraction and reporting. Learn which applications are promising, which require closer scrutiny, and how to decide where AI belongs in your review workflow.
Read Guide →
Artificial Intelligence in Research
Generative AI can classify titles and abstracts against systematic review eligibility criteria, but incorrect exclusions may cause relevant studies to be missed. Learn how AI screening works, how to evaluate its reliability, and what safeguards researchers should use.
Read Guide →
Artificial Intelligence in Research
Generative AI can compare research articles against systematic review eligibility criteria, but reliable classification depends on how criteria are defined and interpreted. Learn why AI may misclassify studies, how to validate its decisions, and when human judgment remains necessary.
Read Guide →
Artificial Intelligence in Research
Generative AI can extract study characteristics and reported results into structured systematic review tables, but errors may affect the evidence synthesis. Learn which information AI can extract, why numerical data require particular care, and how to validate and verify AI-assisted extraction.
Read Guide →
Artificial Intelligence in Research
Generative AI can help locate methodological information and suggest risk-of-bias judgments, but assessing bias requires more than extracting text. Learn how AI can support structured appraisal, why its ratings may be unreliable, and how researchers should verify its conclusions.
Read Guide →
Artificial Intelligence in Research
Generative AI may help identify publications that share participants, recruitment periods, or underlying datasets. Learn how to investigate overlapping study populations, distinguish multiple reports from independent studies, and avoid double-counting evidence in systematic reviews.
Read Guide →
Artificial Intelligence in Research
Overall accuracy can conceal serious AI screening errors, particularly when most search results are irrelevant. Learn how to calculate sensitivity, specificity, precision, and false exclusion rates, interpret a confusion matrix, and determine whether an AI-assisted screening workflow is sufficiently reliable.
Read Guide →
Artificial Intelligence in Research
Reporting AI use in a systematic review requires more than a generic disclosure statement. Learn what to document about AI-assisted screening, data extraction, validation, human oversight, and methodological decisions, and how to distinguish PRISMA reporting from publisher disclosure requirements.
Read Guide →
Artificial Intelligence in Research
Generative AI can suggest theories that may help explain a research problem, but familiar keywords and convincing descriptions do not establish theoretical relevance. Learn how to evaluate AI recommendations using explanatory fit, original sources, and the specific demands of your study.
Read Guide →
Artificial Intelligence in Research
Generative AI can assist researchers throughout a study, but its use may introduce ethical concerns involving participant privacy, informed consent, data handling, discrimination, accountability, and research integrity. Learn how to identify these risks, distinguish research ethics from publication ethics, and determine when AI use requires additional safeguards or institutional review.
Read Guide →
Artificial Intelligence in Research
Generative AI can change the ethical risk of a study when it introduces new data recipients, processing activities, participant interactions, sensitive inferences, or potential consequences for individuals and communities. Learn how to identify meaningful changes in risk and determine when safeguards or institutional review may be necessary.
Read Guide →
Artificial Intelligence in Research
Researchers may need to describe generative AI use in their ethics applications when it affects participant interactions, data processing, confidentiality, or research procedures. Learn what information reviewers need, how to distinguish substantive AI use from routine drafting assistance, and why disclosure alone does not establish ethical approval.
Read Guide →
Artificial Intelligence in Research
Researchers may decide to use generative AI after their study has already received ethics approval. Learn how to determine whether the proposed AI activity is covered by the existing protocol, when an amendment may be necessary, and why consequential changes should be reviewed before implementation.
Read Guide →
Artificial Intelligence in Research
Generative AI may introduce discrimination or stigmatization into research when it assigns unsupported labels, reproduces stereotypes, or interprets participant experiences through biased categories. Learn how to identify these risks, distinguish analytical differences from unfair classifications, and protect participants without assuming that every AI-generated comparison is discriminatory.
Read Guide →
Artificial Intelligence in Research
Generative AI may infer sensitive characteristics from research data even when participants never explicitly provided that information. Learn how these inferences arise, why they may be inaccurate or ethically unjustified, and how researchers can distinguish legitimate analysis from unnecessary profiling.
Read Guide →
Artificial Intelligence in Research
Generative AI may infer participant characteristics from interviews, writing patterns, or contextual information even when those characteristics were not explicitly provided. Learn when such inferences may be methodologically defensible, why plausible predictions are not equivalent to validated measurements, and when researchers should avoid generating them altogether.
Read Guide →
Artificial Intelligence in Research
Introducing generative AI into an approved study may change its expected benefits, potential harms, and the adequacy of existing safeguards. Learn how to reassess the risk-benefit balance without assuming that faster analysis automatically benefits participants or that every AI-related change makes research more dangerous.
Read Guide →
Artificial Intelligence in Research
Evaluating AI-specific ethical risks may require expertise beyond the research team or ethics committee alone. Learn how researchers, research ethics committees, data protection officers, information security specialists, and methodological experts contribute to responsible AI oversight without confusing their distinct responsibilities.
Read Guide →
Artificial Intelligence in Research
Generative AI may offer useful research assistance, but some applications introduce risks that cannot be adequately justified or managed. Learn when avoiding AI is the more defensible decision, why disclosure and technical safeguards are not always sufficient, and how researchers can distinguish appropriate restraint from unnecessary rejection of the technology.
Read Guide →
Artificial Intelligence in Research
Should participants be told when generative AI is used in a research study? Learn when AI involvement is relevant to informed consent, why routine drafting assistance differs from processing participant data, and how researchers can provide meaningful disclosure without making unsupported assurances.
Read Guide →
Artificial Intelligence in Research
Researchers should explain AI processing in enough detail for participants to understand how their information will be used and what risks or choices may matter. Learn how to distinguish meaningful disclosure from unnecessary technical detail, avoid vague assurances, and communicate AI processing arrangements accurately.
Read Guide →
Artificial Intelligence in Research
Research participants may agree to participate in a study while objecting to having their information processed by AI. Learn when an AI-specific opt-out may be appropriate, why it is not universally required, and how researchers can distinguish genuine participant choice from options that the research workflow cannot actually support.
Read Guide →
Artificial Intelligence in Research
Broad consent may permit certain future research uses of participant information, but it does not provide unrestricted permission for every AI-assisted analysis. Learn how the scope of consent, future research purposes, data protection requirements, and institutional oversight affect whether generative AI can be used with previously collected research data.
Read Guide →
