AI across the Research Lifecycle
30-60 minutes
AI tools are reshaping how researchers find literature, write drafts, analyse data, and prepare publications. This experience walks through each stage, with practical activities you can try using your own research.
The goal is not to automate your research, but to identify where AI saves time on routine tasks so you can focus on the thinking that matters.
This experience has 6 steps. Each step covers a different part of the research workflow. Try the activities with your own research topics for the most benefit.
AI tools can help you find relevant papers, summarise abstracts, identify research gaps, and compare methodologies across studies. They work best as a complement to traditional database searches, not a replacement.
Free academic search engine with AI-powered relevance ranking. Good for finding highly cited papers and tracing citation networks. Shows "TLDR" summaries of abstracts.
Ask a research question in plain language and Elicit finds relevant papers, extracts key findings, and organises results into a table. Particularly useful for systematic review screening.
Searches academic papers and synthesises findings into yes/no/mixed answers to your research question. Useful for quickly gauging the state of evidence on a topic.
General-purpose AI can help brainstorm search terms, suggest related concepts, and summarise papers you paste in. Caution: they can hallucinate citations, so always verify references exist.
Pick a research topic you are currently working on. Try these steps:
AI can assist with drafting sections, paraphrasing, adjusting academic tone, and getting structural feedback on your writing. The key is knowing when AI helps and when it starts to erode your distinctive voice.
Give AI your dot points, findings, or rough notes and ask for a structured first draft. This gets words on the page that you then rewrite in your own voice.
Paste a paragraph and ask AI to suggest clearer phrasing, reduce word count, or simplify complex sentences without changing the meaning.
Ask AI to evaluate the logical flow of a section: "Does my argument follow logically? Are there gaps between my evidence and my conclusion?"
If you are writing for a different audience (grant application vs journal article vs conference abstract), AI can help adjust formality and emphasis.
The difference between AI as a shortcut and AI as a thinking partner comes down to how you prompt. Try both of these with the same paragraph from your current draft and compare the results:
You get polished text back, but you have not thought about why your original was unclear. The AI did the thinking for you.
You get diagnostic feedback that makes you a better writer. The AI is a thinking partner, not a ghostwriter.
Notice the difference? The delegation prompt produces text. The conversation prompt produces insight. After trying both, ask yourself: which version helped you understand your own writing better?
AI can assist with coding tasks (R, Python, SPSS syntax), data exploration, pattern spotting, and visualisation. It works best when you understand the analysis conceptually and use AI to speed up the implementation.
Describe what you want to do ("run a mixed-effects model with random intercepts for participant") and AI can generate R or Python code. It can also explain error messages and suggest fixes.
Describe your dataset structure and ask AI to suggest initial exploratory analyses: distributions to check, correlations to examine, potential confounds to consider.
Ask AI to generate ggplot2, matplotlib, or other plotting code. Describe the chart you want and the variables involved, and it will produce publication-ready code.
If you use SPSS, AI can generate syntax for complex procedures, saving you from navigating menus. Describe the analysis in plain language and get syntax you can paste directly.
Describe a dataset you are working with (or a hypothetical one) and try this prompt in ChatGPT or Claude:
Evaluate the suggestions against your own statistical knowledge. Did AI recommend anything inappropriate? Did it surface an approach you had not considered?
Using AI in research comes with disclosure obligations. Journals, funders, and universities increasingly require transparency about how AI was used. Getting this right protects your reputation and your work.
Most major publishers (Elsevier, Springer Nature, Wiley, Taylor & Francis) now require disclosure of AI use in the methods or acknowledgements section. AI cannot be listed as an author. Policies vary by journal, so check before submitting.
ARC and NHMRC have emerging guidance on AI disclosure in grant applications and funded research. Check current guidelines before submitting applications.
Your institution's research integrity framework requires honesty and transparency in all research activities. If AI materially contributed to your work, disclose it. When in doubt, disclose.
Consider: data privacy (did you upload participant data to AI tools?), reproducibility (can others replicate your AI-assisted process?), and bias (does the AI introduce systematic distortions?).
Select a publisher to see their current AI disclosure stance:
Select a publisher above to see their AI disclosure requirements.
You can adapt this template for your own papers:
The AI Toolkit has guides that complement what you have explored here:
More on responsible AI use, data handling, and institutional obligations.
How to work within token limits, structure long conversations, and get better results from complex prompts.
A comprehensive approach to prompting that goes beyond templates, useful for literature synthesis and data analysis tasks.
Practical guidance on protecting research data, choosing enterprise vs free tools, and sustainable AI practices.
Conversation, not delegation. Use AI to challenge your thinking, explore alternatives, and identify gaps, not to produce text you paste in unchanged.
AI accelerates routine research tasks (finding papers, generating code, formatting) but cannot replace the intellectual work of analysis and argumentation.
Always verify AI outputs. Citations, statistical suggestions, and summaries can contain errors that would undermine your work.
Disclose AI use transparently. Check your target journal, funder, and your institution's guidelines before submitting.
Protect research data. Never upload identifiable participant data to public AI tools.
Great work! Your progress has been recorded.
Checking your badge progress...
Once you have worked through the activities above, click below to record your progress.