AI for research: a curated tutorial and video learning guide

Choose a research task, work through one practical resource, and leave with an output you can inspect. This collection pairs original tutorials with selected external guides and recordings. Each recommendation includes a reason to use it and a concrete follow-up activity.

Download the research workflow templates, a ZIP containing editable CSV logs for evidence extraction, citation checking, qualitative coding review, and analysis checks. Open the files in your spreadsheet or research workspace.

You can use the collection independently or as a reading and viewing list for a research group. The external resources remain with their original creators. The descriptions and suggested exercises below are our editorial guidance.

Find papers and organize the evidence

Start with our tutorial: Build an AI-assisted literature review you can verify. Best for a researcher moving from a broad topic to a structured comparison of papers. Your output is a checked matrix with a source location for each extracted claim.

Then consult the tool documentation: Elicit's workflow chooser, by Elicit. It explains the distinction between an agent session, a research report, and a structured systematic-review workflow. Use it to choose an appropriate starting mode; evaluate the resulting evidence yourself. Vendor documentation describes a product's functionality and access conditions.

Keep the reference library tidy: Zotero's duplicate-detection guide, by Zotero. This is useful when you have imported results from several searches. Review a possible duplicate group and check which metadata should be retained before merging.

Check references before writing

Start with our tutorial: How to check AI-generated citations and the claims they support. Best for anyone who has a draft containing AI-suggested references. The output is a claim-by-claim verification log.

Use an independent lookup: Crossref Simple Text Query, by Crossref. It helps match bibliographic references to DOI records. Inspect candidate matches and then read the source to determine whether it supports your sentence.

Try this exercise: choose one paragraph from your own draft. Check every reference and revise any sentence whose scope exceeds the evidence. Record unresolved references separately.

Explore qualitative analysis carefully

Start with our tutorial: Pilot AI-assisted qualitative coding with an audit trail. Best for researchers who already have a qualitative approach and want to evaluate a specific supporting role for AI. It includes a fictional example and a reusable coding instruction.

Read before preparing transcripts: Anonymisation for text data, by the UK Data Service. Its attention to contextual detail is especially useful for interviews and fieldnotes. Use it to review whether apparently anonymous excerpts can still identify someone.

Try this exercise: take an approved or synthetic excerpt, write your own memo, then compare it with the AI's suggestions. Record a useful contribution, an omission, and a point that needs more context.

Learn to question AI-assisted analysis

Start with our tutorial: Check an AI-generated data analysis before using its results. Best for researchers using AI to write R or Python. Your output is a small set of hand-checkable tests and a clean rerun of the analysis.

Watch: Applying generative AI in analytics: failure modes and opportunities, published by INFORMS. This recording belongs beside an analysis tutorial because it puts failure modes on the learning agenda. See our recording page for the event context.

Read alongside it: Research compendia, by the Turing Way Community. Use the guide to organize the data, code, and outputs that make your work inspectable.

Try this exercise: select a transformation in your script, construct a tiny example with a known result, and check the implementation against it.

Discuss responsible research in your group

Watch: Responsible AI in Research: highlights from the NCRM Annual Lecture 2025, published by the National Centre for Research Methods. This is an edited highlights recording. Use it as a discussion starter about where responsibility sits when AI contributes to research.

Watch next: The role of human understanding in the future of scientific discovery, published by Stanford HAI. Pair it with a lab discussion about which judgments your team needs to understand and retain as it adopts automation. Our panel page provides event information.

Try this exercise: choose one current project and write down the AI task, the accountable researcher, the evidence you will retain, and the point at which a human decision is required.

Get an introduction to the initiative

The GAIforResearch introductory webinar introduces the initiative's approach. It was recorded in 2024, so use current documentation for product interfaces and capabilities. The underlying research questions remain useful prompts for discussion.

How to use curated resources well

Check the source, publication date, and fit with your research question. A recording can explain an idea without establishing that a tool works for your dataset. A vendor tutorial can teach a feature without independently evaluating it. Build a small test around the result you need and keep a record of what you learn.

This guide was assembled on 6 November 2026. Linked documentation and the listed YouTube titles and publishing channels were checked on that date. The viewing suggestions are based on the recording metadata and the linked event descriptions, rather than a new full transcript review. Product access and external links can change.

Browse all tutorials, workshops and recordings, or Start here.