RESEARCH & AI TOOLKIT
Research Intelligence Toolkit for every Scholar.
Research tools can accelerate literature discovery, evidence review, citation checking, reference management and analysis. They are valuable when they strengthen scholarly judgement, not when they replace it.
01 · DISCOVER
Use AI search to find evidence, then inspect the evidence itself.
Tools such as Consensus, Semantic Scholar, Google Scholar and ResearchRabbit can reduce the time required to discover relevant literature, but each solves a different problem. Consensus is designed around question-based search over scholarly literature; Semantic Scholar supports broad academic discovery; Google Scholar remains useful for finding versions and citation trails; ResearchRabbit helps researchers move through citation networks rather than relying only on keywords.
The methodological caution is simple: an AI-generated answer is not the same thing as the underlying evidence. A researcher should inspect the studies, populations, methods, dates and limitations behind any summary before incorporating the claim into a dissertation or article.
02 · REVIEW
Elicit and SciSpace can accelerate review tasks, but methodology remains yours.
Elicit supports literature-review activities such as searching, screening, structured extraction and synthesis. SciSpace supports paper discovery, PDF interrogation and literature-review workflows. These capabilities are useful because evidence reviews contain repetitive tasks, but the researcher remains responsible for defining the research question, inclusion criteria and appraisal standards.
For example, if a researcher asks an AI system to identify “relevant studies on digital inclusion,” relevance will depend on how digital inclusion is defined, which populations count, what years are eligible and whether grey literature is included. Those are methodological decisions, not software settings. PRISMA 2020 requires transparent reporting of information sources, search strategies and study selection because credibility depends on how the evidence base was constructed (Page et al., 2021).
03 · VERIFY
A citation count tells you that a paper was cited, not whether later researchers agreed with it.
Scite’s Smart Citations add context by showing whether later citation statements support, contrast with or merely mention a paper. This can help researchers investigate how a claim has been received in subsequent literature. It does not eliminate the need to read the relevant papers, but it is useful for identifying disagreement that ordinary citation counts hide.
Crossref exposes metadata deposited by publishers and other scholarly sources, allowing researchers to check titles, authors, journals, publication dates and DOIs. A source can exist and still be cited incorrectly, so verification should proceed from source existence → metadata accuracy → claim accuracy → later evidence.
04 · REFERENCE & SYNTHESISE
Zotero reduces clerical work; NotebookLM can organise source-grounded thinking.
Zotero supports collection, organisation and citation of research sources and integrates with common word processors. It reduces the mechanical burden of citation management, but metadata should still be checked before submission. NotebookLM can work with user-supplied sources and generate responses grounded in those materials, which can be useful for comparing themes, locating passages and organising a large reading set.
Source-grounded AI is safer than unbounded generation for many research tasks, but it is not error-proof. Important claims should still be checked against the underlying documents.
05 · TOOL DIRECTORY
Choose tools by research task.
