Artificial Intelligence in Research
Research collaboration agreements can clarify which AI tools collaborators may use, how sensitive information is protected, and who remains accountable for AI-assisted work. Learn which provisions matter, when formal clauses are appropriate, and how to avoid unnecessary contractual restrictions.
Read Guide →
Artificial Intelligence in Research
AI can assist with writing, analysis, coding, and research design, but using AI does not automatically establish or eliminate authorship eligibility. Learn how to evaluate human intellectual contributions, apply disciplinary authorship criteria, and preserve accountability in AI-assisted research.
Read Guide →
Artificial Intelligence in Research
AI can generate manuscript text, code, analytical recommendations, and research interpretations, making it harder to determine which intellectual contributions came from human collaborators. Learn how to distinguish tool use from substantive scholarly work and document contributions fairly.
Read Guide →
Artificial Intelligence in Research
AI assistance does not transfer scholarly responsibility away from human authors. Learn how research teams can preserve collective accountability through defined roles, qualified verification, transparent communication, and meaningful coauthor approval.
Read Guide →
Artificial Intelligence in Research
AI errors in high-stakes research can affect participant welfare, clinical decisions, legal rights, and consequential scientific findings. Learn how researchers can adapt AI use through risk-proportionate verification, qualified oversight, and appropriate restrictions.
Read Guide →
Artificial Intelligence in Research
An AI-assisted research task becomes high stakes when errors could cause serious consequences for participants, research validity, legal rights, or consequential decisions. Learn how to assess risk based on context rather than assuming every use of AI carries the same implications.
Read Guide →
Artificial Intelligence in Research
AI may support clinical research decisions, but its recommendations require safeguards proportionate to their consequences. Examine the distinction between assistance and autonomous decision-making, the evidence needed for validation, and the responsibilities of clinical researchers.
Read Guide →
Artificial Intelligence in Research
AI can help researchers identify potentially eligible participants, but incorrect classifications may affect safety, fairness, and research validity. Learn when AI-assisted eligibility screening may be appropriate and why final decisions require authorized review.
Read Guide →
Artificial Intelligence in Research
AI may help researchers identify potential participant safety concerns, but missed risks and incorrect alerts can have serious consequences. Learn how to evaluate AI-assisted safety monitoring, validate alert systems, preserve qualified human judgment, and protect participants.
Read Guide →
Artificial Intelligence in Research
AI may help researchers organize complex legal or ethical information, but generated interpretations can omit context, misstate obligations, or expose confidential data. Learn when AI assistance may be appropriate and why consequential interpretations require qualified human review.
Read Guide →
Artificial Intelligence in Research
AI may support research involving children and other vulnerable populations, but additional safeguards are often necessary. Learn how to evaluate consent, privacy, participant welfare, and the limits of AI-assisted decisions.
Read Guide →
Artificial Intelligence in Research
AI can assist with analyzing sensitive participant information, but inaccurate classifications, privacy breaches, and stigmatizing interpretations may cause substantial harm. Understand when such analysis is justified and which safeguards researchers should establish.
Read Guide →
Artificial Intelligence in Research
AI errors can have consequences beyond inaccurate research findings, particularly when they affect participant safety, rights, or access to services. Learn how to evaluate potential harm, establish appropriate safeguards, and determine when AI should not be used.
Read Guide →
Artificial Intelligence in Research
Human review is essential in many high-stakes AI research applications, but simply having a person approve an output may not provide adequate protection. Learn how to determine the necessary depth, independence, and frequency of verification based on the consequences of error.
Read Guide →
Artificial Intelligence in Research
Double human verification can strengthen oversight of consequential AI-assisted research decisions, but two reviewers do not automatically guarantee accuracy. Learn when independent review is warranted, how to structure it, and when other safeguards may be more effective.
Read Guide →
Artificial Intelligence in Research
Some AI applications should not proceed, even when the technology appears capable or efficient. Learn how to recognize unacceptable consequences of error, identify situations where safeguards are insufficient, and select more defensible research methods.
Read Guide →
Artificial Intelligence in Research
AI can help identify observations that may require exclusion, but it should not independently determine which data belong in a final analysis. Learn how to establish defensible exclusion criteria, investigate flagged observations, and document decisions without introducing analytical bias.
Read Guide →
Artificial Intelligence in Research
AI may help detect safety concerns and analyze evidence relevant to stopping a research study, but decisions to suspend or terminate research require appropriate authority and scientific justification. Learn how stopping rules, interim analyses, participant protection, and independent oversight should guide these decisions.
Read Guide →
Artificial Intelligence in Research
AI can recommend analytical methods, exclusions, statistical models, and interpretations, but researchers must remain accountable for consequential decisions. Learn why human attribution matters, how to document AI-assisted analysis, and what meaningful statistical oversight requires.
Read Guide →
Artificial Intelligence in Research
AI can help identify suspicious patterns in manuscripts, datasets, images, and publication records, but a detection flag is not proof of research misconduct. Learn what AI can detect, where its reliability is limited, and how researchers and institutions should evaluate potential integrity concerns.
Read Guide →
Artificial Intelligence in Research
AI can identify suspicious numerical patterns, inconsistencies, and possible signs of fabricated research data, but no statistical detector can reliably prove fabrication on its own. Learn how to investigate anomalies, assess data provenance, and distinguish genuine observations from manufactured evidence.
Read Guide →
Artificial Intelligence in Research
AI can help identify duplicated regions, suspicious similarities, and possible alterations in scientific images, but an automated flag cannot establish falsification or intent. Learn how image-forensic tools work, what their limitations are, and how researchers should verify concerns using original data and appropriate integrity procedures.
Read Guide →
Artificial Intelligence in Research
AI can help identify patterns associated with paper-mill manuscripts, including repeated templates, suspicious images, questionable references, and coordinated submission activity. However, these indicators do not independently establish publication fraud. Learn how researchers and editors can evaluate potential paper-mill submissions using verifiable evidence and fair review procedures.
Read Guide →
Artificial Intelligence in Research
AI can help identify unusual reviewer identities, suspicious review patterns, and possible manipulation of editorial processes, but it cannot independently establish that a peer review is fraudulent. Learn how editors can investigate suspicious reviews, verify reviewer identities, and protect the integrity of scholarly evaluation.
Read Guide →
Artificial Intelligence in Research
AI research-integrity tools can incorrectly flag legitimate manuscripts, datasets, and scientific images as suspicious. Learn why false positives occur, how detection accuracy differs from evidentiary reliability, and what researchers, editors, and institutions should do to prevent unsupported misconduct allegations.
Read Guide →
