AI tools for PhD students need to handle the long work: reading a whole field, tracing citations, reviewing chapters of 80,000 words and getting ready to defend them. AI for Research puts Deep and Detailed literature reports, a Citation Map, a Systematic Review workspace, Data Lab, a PhD-level Reviewer and defense preparation in one place, with every answer citing real papers.
Three jobs you can do here
1. Read and map your field
A Deep report (15 credits) answers a question in 2,500 to 4,000 words with figures, from about 250 papers screened and 12 read in full. A Detailed report (45 credits, paid plans) runs 5,000 to 8,000 words with a study table from 90+ papers, and shows you its plan of sub-questions to edit before it starts. Then open the key papers in the Citation Map (free): references and citing papers as a graph, snowballing rounds across your set, and the reviews and meta-analyses that cite it.
2. Get each chapter reviewed at PhD level
Upload a chapter to the Reviewer, choose Thesis and PhD, and run the checks you need: structure and argument, method and statistics, citations, clarity, formatting. Each criterion is scored out of 10, and every comment quotes your passage and says how to fix it. Standard costs 10 credits per 10,000 words; High quality (Pro) costs 30 and adds score charts and issues by section.
3. Prepare the defense and publish from the thesis
Defence preparation in the Thesis Suite (3 credits) reads your thesis or abstract and lists likely examiner questions chapter by chapter, with how to answer them. Slides builds the defense deck with the thesis-defense preset (10 credits). Thesis to manuscript (10 credits) turns a chapter into a journal article draft.
How it works, step by step
- Set up a project for the thesis and invite your supervisor as Can comment and any co-author as Can edit.
- Import your library from Zotero or Mendeley (paid plans), or upload PDFs.
- Ask your chapter's question at Deep or Detailed depth. Choose sources (All papers, Medical, Preprints, Web, My Library) and filters such as year and study type.
- Snowball from the strongest papers in the Citation Map; export RIS or CSV to your reference manager.
- If your thesis has a systematic review, run it in the Systematic Review workspace: protocol, logged searches, screening with a second reviewer, extraction, PRISMA 2020 diagram.
- Analyze data in Data Lab: t-tests, ANOVA, chi-square, correlation and regression run free in the browser, and the AI results section (2 credits) is written from the computed numbers only.
- Review each chapter at PhD level, then run Defence preparation on the full thesis.
The workspaces you will use
| Workspace | What you do with it | Credits |
|---|---|---|
| Deep Research | Summary, Deep and Detailed reports | 1, 15, 45 (High quality ×3, Pro) |
| Citation Map | References, citing papers, snowballing | 0 |
| Systematic Review | Protocol to PRISMA diagram | AI screening 1 per 25 records |
| Data Lab | Statistics and charts; AI results section | 0; write-up 2 |
| Reviewer | PhD-level review of chapters | 10 or 30 per 10,000 words |
| Thesis defense prep | Examiner questions by chapter | 3 |
| Research slides | Defense deck, PowerPoint with speaker notes | 10 |
| Thesis to manuscript | Chapter to journal article draft | 10 |
Which plan fits, and a month worked out
Pro fits a PhD in its writing-up years: 1,500 credits a month, High quality mode, 40 Deep or Detailed reports a day with 3 at once, unlimited projects with 10 people each, PDFs up to 100 MB. It costs $29 a month or $228 a year. Plus ($12, 750 credits, Detailed reports, one report at a time) is enough in reading years. For the final stretch, the Defence Pass is $49 once for 90 days of Pro.
Example month (invented, on Pro): year three, a public-health PhD with a systematic review chapter.
| Job | How many | Credits |
|---|---|---|
| Detailed reports | 2 | 90 |
| Deep reports | 6 | 90 |
| Summary answers | 40 | 40 |
| Questions to PDFs | 80 | 80 |
| AI screening suggestions, 1,200 records | 48 batches of 25 | 48 |
| AI extraction, 40 included studies | 4 batches of 10 | 12 |
| Protocol draft and Methods draft | 2 | 6 |
| Data Lab results sections | 3 | 6 |
| Reviewer, High quality, a 25,000-word chapter set | 3 units of 10,000 words | 90 |
| Defence preparation runs | 5 | 15 |
| Thesis to manuscript | 2 | 20 |
| Defense slide deck | 1 | 10 |
| Citation Map, Landscape, statistics, PRISMA diagram | any | 0 |
| Total | 507 of 1,500 |
The High quality review is the line that needs Pro; the other 417 credits would fit on Plus.
AI-use rules that matter to you
Two sets of rules apply: your university's for the thesis, and each journal's for the papers that come out of it.
- Thesis. Read your graduate school's regulations on AI, agree uses with your supervisor in writing, and keep a log to write the declaration from.
- Journal papers. ICMJE says chatbots cannot be authors, authors must disclose AI use and are responsible for all content, and AI output cannot be cited as a primary source [1]. COPE says AI tools cannot be authors and authors must disclose how AI was used [2]. Nature Portfolio asks for AI use to be documented, usually in the Methods, and does not accept generative AI images except where AI is the research subject [3].
- Reviewing for journals. If a journal sends you a manuscript to review, do not upload it to any AI tool, including this one; Nature Portfolio asks reviewers not to [3].
What it does not do
- No Google Scholar, Scopus or Web of Science search; those come in only as RIS or CSV exports to a systematic review.
- No meta-analysis or forest plot yet; Data Lab has no Wilcoxon, Kruskal-Wallis or logistic regression. Run those in R, jamovi or SPSS.
- Reports read open full texts; paywalled papers are read from the abstract.
- The Reviewer and Defence preparation do not predict your examiners.
Who else it is for
On a master's? See AI for a master's thesis. After the PhD, see AI for professors.
Publishing during the doctorate? See the conference paper writer, the manuscript formatting tool and the research poster maker. For analysis, see qualitative coding and the Bayesian t-test calculator. The sources for your discipline are on the field pages, for example life sciences and engineering and computing.
Learn more
Sources
- International Committee of Medical Journal Editors. Recommendations: AI use by authors. https://www.icmje.org/recommendations/browse/artificial-intelligence/ai-use-by-authors.html
- Committee on Publication Ethics. Authorship and AI tools: COPE position statement (13 February 2023). https://publicationethics.org/cope-position-statements/ai-author
- Nature Portfolio. Editorial policies: Artificial intelligence (AI). https://www.nature.com/nature-portfolio/editorial-policies/ai
Last updated 2026-11-13