AI in research: an introduction
▶️ Watch: Tony Myers on AI in research
Artificial intelligence can support academic research when its role is tightly defined. It can help organise material, generate search terms, critique drafts, draft code, and reformat outputs for different audiences. It should not be used as a citation source or as the final authority on evidence.
From literature mapping and grant preparation through to analysis support and dissemination, AI is most useful when paired with clear prompts, secure platform choices and independent verification. Research databases, primary sources, reproducible code and expert judgement still carry the evidential weight.
Full research guide
The research subsite expands this page with secure/private platform choices, a seven-part research prompt, the VALID-AI checklist, source-grounded NotebookLM/RAG guidance, and verification protocols.
Open the Research GuideAI is a collaborator, not evidence
Use AI for generation, critique, organisation and transformation. Use scholarly sources, data and expert judgement for evidence.
Verification is structural
Build checks into prompts and workflows instead of treating verification as a final tidy-up.
Source-grounded is not magic
NotebookLM and RAG can help interrogate a curated source set, but citations and interpretations still need checking.
The research lifecycle
AI techniques by research stage
Open a stage to explore techniques, prompt examples and tools for that part of the research process.
🔎 Literature discovery & synthesis
Use AI to move beyond keyword searches into a conceptual exploration of knowledge. Tools can map citation networks to find seminal works or synthesise findings from hundreds of papers.
Practical applications
- Synthesise key themes and arguments from a large collection of research papers.
- Build visual citation maps to discover seminal and derivative works.
- Automate the extraction of data for systematic literature reviews into structured tables.
- Generate summaries of articles to quickly assess their relevance.
Prompt examples
Tools
Research Rabbit · Elicit · Connected Papers · Semantic Scholar
📊 Qualitative & quantitative analysis
Augment data analysis by using AI to perform preliminary coding on qualitative data (transcripts, field notes) or to help write and debug code for quantitative statistical analysis.
Practical applications
- Generate Python or R code for complex statistical tests and data visualisations.
- Perform preliminary thematic coding of interview transcripts to identify initial patterns.
- Clean and pre-process datasets by identifying anomalies or missing values.
- Interact with spreadsheets using natural language to perform quick calculations and create charts.
Prompt examples
Tools
✍️ Grant proposals & funding
Leverage AI to streamline the grant writing process. Use it to identify suitable funding bodies, structure proposals, and articulate the impact of your research.
Practical applications
- Summarise the key requirements and priorities from a funding call document.
- Draft specific sections of a grant proposal, such as the background, literature review, or impact statement.
- Reframe your research summary to align with the strategic objectives of different funders.
- Get feedback on the clarity and persuasiveness of your proposal narrative.
Prompt examples
Tools
ChatGPT · Claude · NotebookLM · Elicit
📝 Writing & productivity
Ethically use AI as a writing assistant for brainstorming, outlining, and proofreading. Specialised tools can also help check your argument structure and suggest citations.
Practical applications
- Improve the clarity, flow, and conciseness of manuscript drafts.
- Receive feedback on the logical structure and coherence of an argument.
- Summarise your article for different purposes (e.g., abstract, lay summary, blog post).
- Format citations and bibliographies according to specific style guides.
Prompt examples
Tools
Thesify · NotebookLM · Grammarly · ChatGPT
📢 Dissemination & engagement
Amplify the reach of your research by using AI to translate complex findings into accessible formats for diverse audiences.
Practical applications
- Create a slide deck outline for an academic conference presentation.
- Write plain-language summaries of your research for policy briefs or press releases.
- Generate social media posts (e.g., for X, LinkedIn) to share key findings with a wider audience.
- Create scripts for short educational videos or podcasts explaining your work.
Prompt examples
Tools
For secure platform choices, the VALID-AI checklist, verification protocols and disclosure templates, continue to the full research guide.
Companion guides