If you're learning about systematic reviews, it probably feels like just a more detailed literature review. But that framing is misleading. Discover...
Explore Playlists
Latest Posts
Do you find terms like ontology and epistemology alien and confusing? You are not alone! Use our simple "iceberg analogy" to understand the core...
Struggling with your literature review? Discover a simple, 6-step process to search, evaluate, organize, and write a literature review for your...
Distracted while writing your dissertation or academic report? Put your headphones on and use this 3-hour, non-stop concentration music track to...
Continually embrace sustainable benefits with revolutionary metrics. Completely benchmark premier partnerships through multifunctional convergence...
Designing a protocol for a semi-structured interview doesn't have to be confusing. Discover 6 straightforward tips to structure your questions, avoid...
Frequently Asked Questions
Responsible AI in Evidence Synthesis
Clear answers on using AI for screening, extraction and synthesis without compromising rigour, transparency, or trust.
A Section A Foundations 5 Questions
1What is responsible AI in evidence synthesis?
Responsible AI in evidence synthesis is the use of AI systems in ways that preserve scientific rigour, transparency, accountability, reproducibility and trustworthiness. AI should support evidence synthesis without compromising research integrity.
2Why is responsible AI important?
Evidence syntheses inform health, policy and practice decisions. If AI introduces errors, bias or opaque decision-making, trust in the evidence may be undermined.
3Why is AI being introduced into evidence synthesis?
The volume of research is growing faster than humans can process it. AI may help reduce workload, improve efficiency and support more timely evidence synthesis.
4What problems can AI help solve?
AI can help with searching, screening, data extraction, summarisation, translation and maintaining living evidence systems.
5What problems can AI create?
AI can introduce hallucinations, bias, opacity, privacy risks, copyright concerns and reproducibility challenges.
B Section B Accountability 5 Questions
6Who is responsible for AI-assisted evidence synthesis?
The evidence synthesis team remains responsible for all methods, decisions and conclusions, regardless of which AI tools are used.
7Can responsibility be delegated to AI?
No. AI may assist with tasks, but accountability remains human.
8Can AI be listed as an author?
No. AI systems cannot meet authorship criteria and should not be credited as authors.
9What is human oversight?
Human oversight means that people supervise AI outputs, verify important decisions and remain accountable for outcomes.
10Why is human oversight necessary?
Because AI can make mistakes, produce misleading outputs and fail in unexpected ways.
C Section C AI Literacy 3 Questions
11What is AI literacy?
AI literacy is the ability to understand, evaluate and use AI systems responsibly.
12How much technical knowledge do evidence synthesists need?
Evidence synthesists do not need to be AI engineers, but they should understand how tools work, how they were evaluated and their limitations.
13Why does AI literacy matter?
Without AI literacy, users may overtrust AI systems or misuse them.
D Section D Tool Selection 5 Questions
14How should I decide whether to use an AI tool?
Use AI only when there is evidence that the tool is suitable for the task and can be used without compromising methodological rigour.
15Does the existence of AI justify its use?
No. AI use should always be justified based on evidence and context.
16What evidence should support adoption of an AI tool?
Validation studies, benchmarking results, independent evaluations and evidence from similar review contexts.
17Can a validated tool be unsuitable?
Yes. Validation in one context does not guarantee suitability in another.
18What should organisations consider before adopting AI?
Governance, privacy, validation evidence, cost, training requirements and ongoing monitoring.
E Section E Validation 6 Questions
19What is validation?
Validation is the process of testing whether an AI system performs reliably for its intended task.
20Why is validation important?
Without validation, there is no evidence that a tool can be trusted.
21What is generalisability?
Generalisability is whether a tool performs reliably outside the setting where it was originally evaluated.
22What is robustness?
Robustness is the ability of a tool to perform well when confronted with unusual or difficult inputs.
23What is stability?
Stability refers to whether a tool produces consistent outputs when given the same inputs repeatedly.
24What is a SWAR?
A Study Within A Review is an evaluation embedded within a real review to assess AI performance.
F Section F AI Tasks 6 Questions
25Can AI search for evidence?
AI can support searching, but search remains a high-risk task that requires careful validation.
26Can AI screen studies?
Yes. Screening is one of the most promising applications of AI in evidence synthesis.
27Can AI extract study data?
Yes, particularly when human verification remains in place.
28Can AI assess risk of bias?
AI may assist, but risk of bias assessment remains heavily dependent on human judgement.
29Can AI perform evidence synthesis?
AI can assist with summarisation, but evidence synthesis still requires human methodological judgement.
30Can AI perform meta-analysis?
Current generative AI systems cannot independently perform a valid meta-analysis.
G Section G Summarisation vs Synthesis 4 Questions
31Is summarising studies the same as evidence synthesis?
No. Evidence synthesis involves methodological processes that go beyond summarisation.
32Why is evidence synthesis more than summarisation?
Because synthesists assess quality, bias, heterogeneity, certainty and relevance before drawing conclusions.
33What is decision-grade evidence?
Evidence that has been produced using rigorous and transparent methods and is suitable to inform decisions.
34Can a chatbot provide decision-grade evidence?
Not on its own. Decision-grade evidence requires systematic methods and human oversight.
H Section H Hallucinations & Errors 4 Questions
35What is a hallucination?
A hallucination is information generated by AI that is unsupported, fabricated or incorrect.
36Why do hallucinations matter?
Because they may introduce false information into evidence syntheses.
37Should AI-generated citations be checked?
Always.
38Can hallucinations be eliminated?
No. They can be reduced but not completely eliminated.
I Section I Bias & Fairness 3 Questions
39What is algorithmic bias?
Systematic error introduced by training data, model design or implementation choices.
40Can AI amplify bias?
Yes. AI may reproduce or magnify biases already present in research systems.
41How can bias be reduced?
Through diverse datasets, validation, transparency and human oversight.
J Section J Transparency & Reporting 4 Questions
42What AI use should be reported?
Any use that influences review methods, decisions, findings or outputs.
43What information should be reported?
Tool name, version, purpose, workflow stage, prompts where appropriate and verification processes.
44What is provenance?
A record showing where information came from and how it was generated or modified.
45Why is provenance important?
It supports transparency, accountability and reproducibility.
K Section K Ethics & Governance 4 Questions
46What ethical issues arise from AI use?
Bias, fairness, privacy, transparency, environmental impact and responsible use of training data.
47What copyright issues arise from AI use?
Researchers must ensure they have appropriate rights to upload, process and share copyrighted materials.
48What privacy issues arise from AI use?
Sensitive or confidential information may be exposed if appropriate safeguards are not in place.
49What is responsible handover?
The process of ensuring that users understand how an AI tool works, its limitations and how it should be used safely.
L Section L The Future 1 Question
50What is the most important principle for responsible AI in evidence synthesis?
AI should only be used when it improves efficiency or capability without compromising rigour, transparency, accountability, reproducibility or trust.
No questions match your search. Try a different term.