Write a Research Paper with AI Assistance

1. 1

   Refine Your Research Question and Scope

   A good research paper starts with a precise, answerable question. AI can help you move from a broad topic to a focused research question by mapping the existing literature and identifying genuine gaps.

   Claude ChatGPT Elicit

   I'm writing a research paper and need help refining my research question.
   
   **My field**: [e.g., computer science, psychology, economics, education, environmental science]
   **Broad topic area**: [e.g., 'the effect of social media on teenagers' mental health']
   **My level**: [undergraduate / master's / PhD / postdoc / independent researcher]
   **Paper type**: [course assignment / thesis chapter / journal submission / conference paper]
   **Target journal/conference** (if applicable): [name, so we can match scope and style]
   **What I already know**: [summarize your current understanding in 3-5 sentences]
   **Data I have or can access**: [e.g., 'I have survey data from 200 participants' / 'I'll use publicly available datasets' / 'theoretical paper, no empirical data']
   
   Help me refine my research question:
   
   1. **Current Question Assessment**: Is my topic too broad, too narrow, or about right for a [paper type]? Be honest.
   
   2. **Research Question Options**: Generate 5 specific, focused research questions within my broad topic. For each:
      - The question itself (in standard academic phrasing)
      - Why this question matters (significance)
      - Whether it's likely answerable with my available data/methods
      - How it differs from what's already been studied
      - Potential methodology to answer it
      - Difficulty level (given my academic level)
   
   3. **Gap Identification**: Based on your knowledge of this field, what are 3 genuine gaps in the existing research that my paper could address? Don't give me gaps that have already been filled — I'll verify these against real literature.
   
   4. **Scope Calibration**: For my chosen question:
      - What should definitely be IN scope?
      - What should explicitly be OUT of scope (to keep the paper focused)?
      - What are the key variables or concepts I need to define?
      - What theoretical framework(s) should I use?
   
   5. **Working Thesis/Hypothesis**: Based on the research question I choose, help me draft a preliminary thesis statement or set of hypotheses.

   Copy

   Tip: Run your research question through this test: Can you clearly identify the independent variable, the dependent variable, and the population? If not, your question is too vague. 'How does X affect Y in Z population under W conditions?' is specific enough. 'What are the impacts of technology on learning?' is not.

2. 2

   Conduct AI-Assisted Literature Review

   Use AI to accelerate literature review -- finding relevant papers, summarizing them, identifying themes, and spotting gaps. This is where AI saves the most time: what takes days manually takes hours with AI assistance.

   Elicit Consensus Scispace Claude

   Help me conduct a systematic literature review for my research paper.
   
   Research question: [your refined question from Step 1]
   Field: [your field]
   Time scope: [e.g., papers from the last 10 years, or 'seminal works from any period + recent advances']
   
   **Phase 1 — Paper Discovery:**
   Based on my research question, suggest:
   - 15-20 key search terms and boolean combinations to use in Google Scholar, Scopus, and Web of Science
   - 5 seminal/foundational papers I must include (the papers everyone in this area cites)
   - 5 recent papers (last 2-3 years) that represent the current state of the field
   - 3 key review/meta-analysis papers that summarize the field
   
   **Phase 2 — Paper Summarization:**
   For each paper I provide (I'll paste abstracts or full text), create a structured summary:
   
   | Field | Content |
   |-------|----------|
   | Authors & Year | |
   | Research Question | |
   | Methodology | |
   | Key Findings | |
   | Sample Size/Data | |
   | Limitations noted | |
   | How it relates to MY research | |
   | Key quotes to cite | |
   | Strengths I can build on | |
   | Gaps MY paper can fill | |
   
   **Phase 3 — Thematic Synthesis:**
   After I've fed you 15-20 paper summaries, organize the literature into:
   - Major themes/schools of thought (with papers sorted into each)
   - Points of consensus (what most researchers agree on)
   - Points of debate (where researchers disagree — this is where my paper adds value)
   - Methodological approaches used and their relative strengths
   - Chronological evolution of thinking on this topic
   - A clear statement of the research gap my paper addresses
   
   **Phase 4 — Literature Review Draft:**
   Write a structured literature review (target: [word count, e.g., 2000 words]) organized thematically (not paper-by-paper). Each paragraph should:
   - Start with a claim or theme
   - Support it with citations from multiple papers
   - Identify tensions or gaps
   - Build toward my research question's justification
   
   Use proper academic citation format: [APA 7th / Chicago / Harvard / IEEE / specify yours]

   Copy

   Tip: NEVER trust AI-generated citations without verification. AI frequently invents plausible-sounding papers with real-sounding author names that don't exist. Use AI to help you discover and summarize real papers, but always verify every citation in Google Scholar, Scopus, or your university library. A fabricated citation in a published paper is academic career suicide.

3. 3

   Build Your Paper Outline and Argument Structure

   Create a detailed outline that maps your argument from introduction to conclusion, with each section building logically on the previous one. A strong outline makes the actual writing almost mechanical.

   Claude ChatGPT

   Create a detailed outline for my research paper.
   
   Research question: [your question]
   Thesis/hypothesis: [your thesis]
   Methodology: [briefly describe your approach]
   Key findings (if you have data already): [summarize]
   Target word count: [e.g., 6,000 words for a journal article, 10,000 for a thesis chapter]
   Citation style: [APA / Chicago / IEEE / other]
   
   Outline with these sections:
   
   **1. Introduction** (target: [X] words):
   - Hook: What makes this topic important RIGHT NOW?
   - Background: What does the reader need to know? (3-4 sentences of context)
   - Problem statement: What gap exists in current knowledge?
   - Research question(s): Clearly stated
   - Significance: Why does answering this question matter? (for the field AND for practice)
   - Paper structure: Brief roadmap of what follows
   
   **2. Literature Review** (target: [X] words):
   - Theme 1: [from your synthesis in Step 2]
     - Sub-theme A with key citations
     - Sub-theme B with key citations
   - Theme 2: [next theme]
     - Sub-themes with citations
   - Theme 3: [next theme]
   - Research gap summary: How the reviewed literature points to my specific question
   - Theoretical framework: [your framework] and why it's appropriate
   
   **3. Methodology** (target: [X] words):
   - Research design: [qualitative / quantitative / mixed methods]
   - Data collection: [how, when, from whom, how much]
   - Sample/participants: [who, how selected, why this sample]
   - Variables and measures: [what you measured, what instruments/tools]
   - Data analysis approach: [statistical methods, coding approach, etc.]
   - Validity and reliability: How you ensured rigor
   - Ethical considerations: IRB approval, consent, data privacy
   - Limitations of the methodology (be upfront)
   
   **4. Results/Findings** (target: [X] words):
   - Organize by research question or hypothesis
   - For each finding: state it, present the evidence, note the significance level
   - Tables and figures: what data should be visualized?
   
   **5. Discussion** (target: [X] words):
   - Interpretation of findings in context of existing literature
   - How do findings align with or challenge previous research?
   - Theoretical implications
   - Practical implications
   - Limitations of THIS study
   - Future research directions
   
   **6. Conclusion** (target: [X] words):
   - Restate the core finding (not a copy-paste of the abstract)
   - The 'so what?' — why this matters
   - Call to action for future researchers or practitioners
   
   For each section and sub-section, specify:
   - The main argument/point
   - Key evidence or citations to include
   - Connection to the next section (logical flow)
   - Approximate word count

   Copy

   Tip: Your Discussion section is where papers succeed or fail. Many writers just restate their findings — but the Discussion should INTERPRET them. Ask: How do my findings change what we thought we knew? What surprised me, and what does that surprise tell us? What should someone do differently because of this research? If your Discussion could apply to any study, it's too generic.

4. 4

   Draft, Revise, and Polish the Paper

   Write the paper section by section with AI assistance, then do multiple revision passes focusing on different quality dimensions: argument strength, clarity, academic tone, and technical accuracy.

   Claude ChatGPT Grammarly

   Help me draft and revise Section [number/name] of my paper.
   
   [Paste your outline for this section]
   [Paste any notes, data, or preliminary writing you have for this section]
   
   **Drafting Guidelines:**
   - Academic writing standard: [journal/conference level / thesis / coursework]
   - Tone: [formal academic, but clear and readable — avoid unnecessary jargon]
   - Person: [first person plural 'we' / third person / passive voice — match your field's convention]
   - Citation style: [your style]
   - Key points this section must make: [list them in priority order]
   
   **After drafting, run these revision checks:**
   
   1. **Argument Strength**:
      - Does every paragraph have a clear topic sentence?
      - Does each claim have sufficient evidence or citation?
      - Are there any logical leaps or unsupported assertions?
      - Does the argument flow logically from one paragraph to the next?
      - Flag any sections where the reasoning is weak or circular
   
   2. **Academic Writing Quality**:
      - Replace vague language with precise language ('many researchers' → cite them or specify a number)
      - Eliminate hedging where the evidence is strong ('may suggest' → 'indicates' when appropriate)
      - Add hedging where the evidence is uncertain (avoid overclaiming)
      - Check for consistency in terminology (don't switch between synonyms for key concepts)
      - Remove colloquialisms that don't belong in academic writing
   
   3. **Clarity Pass**:
      - Flag any sentence longer than 35 words (break it up)
      - Flag any paragraph longer than 6 sentences (split it)
      - Identify jargon that needs definition on first use
      - Ensure acronyms are defined on first use
   
   4. **Citation Check**:
      - Are all factual claims cited?
      - Is any section over-reliant on a single source?
      - Are there places where additional citations would strengthen the argument?
      - Are citations in the correct format?
   
   5. **Section Transition**:
      - Write a transition from this section to the next
      - Does the reader know why they're moving to the next topic?
   
   Return the full drafted section with revision notes as inline comments.

   Copy

   Tip: Write your Introduction and Conclusion LAST, even though they come first in the paper. You can't properly introduce a paper you haven't finished writing, and you can't conclude arguments you haven't fully developed. Draft the methods and results first, then the discussion, then the introduction and conclusion.

5. 5

   Abstract, Formatting, and Submission Preparation

   Write the abstract, format everything to journal requirements, do a final quality check, and prepare supplementary materials. The abstract is the most-read part of any paper — it deserves its own dedicated step.

   Claude ChatGPT

   My paper is drafted and revised. Help me prepare it for submission.
   
   **1. Abstract** (write 3 versions, I'll pick the best):
   Target length: [e.g., 250 words for most journals]
   Structured or unstructured: [some journals require Purpose/Methods/Results/Conclusions structure]
   
   Requirements for each version:
   - One sentence per element: background, gap, method, key finding(s), implication
   - Include the primary keyword(s): [list them]
   - Make it self-contained: someone should understand the paper's contribution from the abstract alone
   - End with a statement of significance, not just a summary of findings
   - Version A: Emphasis on findings
   - Version B: Emphasis on methodology novelty
   - Version C: Emphasis on practical implications
   
   **2. Keywords**:
   - Suggest 5-7 keywords for indexing
   - Mix of broad terms (for discoverability) and specific terms (for targeted searches)
   - Include at least one keyword not in the title
   
   **3. Formatting Checklist** for [target journal/conference]:
   - [ ] Title page formatted correctly (title, authors, affiliations, corresponding author, word count)
   - [ ] Abstract within word limit
   - [ ] Headings match journal style (numbered? Bold? Capitalization?)
   - [ ] Citations in correct format throughout
   - [ ] Reference list in correct format and order
   - [ ] Tables formatted to journal spec (no vertical lines? APA table style?)
   - [ ] Figures at required DPI and format
   - [ ] Line spacing and margins correct
   - [ ] Page numbers present
   - [ ] Supplementary materials properly labeled and referenced
   
   **4. Pre-Submission Quality Check**:
   Review the full paper for:
   - Any remaining [TODO], [CITE], or [CHECK] placeholders
   - Consistency: do numbers in the abstract match numbers in the results?
   - Are all figures and tables referenced in the text?
   - Does the paper answer the research question stated in the introduction?
   - Is the contribution clearly stated and supported?
   
   **5. Cover Letter**:
   Draft a cover letter to the journal editor:
   - Why this paper fits this journal
   - The paper's main contribution in 2-3 sentences
   - Confirmation it's not under review elsewhere
   - Suggested reviewers (if required): recommend 3 researchers in the field who could review this paper
   
   **6. AI Usage Disclosure**:
   Based on current academic norms and [target journal]'s policy:
   - What should I disclose about AI use in writing this paper?
   - Where should the disclosure go? (acknowledgments, methods, author statement)
   - Draft an appropriate disclosure statement

   Copy

   Tip: Run your abstract through this test: send it to someone outside your field and ask them to tell you what the paper found and why it matters. If they can't answer both questions clearly after reading your abstract, it needs revision. Your abstract will be read 100x more than your full paper — it's your most important marketing material.

6. 6

   Peer Feedback Simulation and Response Preparation

   Before submitting, simulate the peer review process. AI can anticipate likely criticisms and help you either fix them preemptively or prepare responses.

   Claude ChatGPT

   Act as a rigorous peer reviewer for my paper. I'm submitting to [journal/conference name] in the field of [your field].
   
   Here's my full paper (or key sections):
   [Paste paper or key sections]
   
   Review as if you are:
   - Reviewer 1: An expert in [your specific topic area] who focuses on methodological rigor
   - Reviewer 2: A senior researcher in [broader field] who focuses on contribution and significance
   - Reviewer 3: A methods specialist who will scrutinize your statistical analysis / qualitative coding / experimental design
   
   For each reviewer persona, provide:
   
   1. **Overall Assessment**: Accept / Minor Revisions / Major Revisions / Reject
      With reasoning.
   
   2. **Major Concerns** (2-3 per reviewer):
      - What's wrong and why it matters
      - What they'd need to see to be convinced
      - Difficulty to address: easy fix vs. requires new data/analysis
   
   3. **Minor Concerns** (3-5 per reviewer):
      - Specific line-level or paragraph-level issues
      - Suggestions for improvement
   
   4. **Questions They'd Ask**:
      - 5 tough questions the reviewers are likely to raise
      - For each, draft a response I could include in a revision letter
   
   5. **Pre-emptive Fixes**:
      Based on the simulated review, list 5 changes I should make BEFORE submission to preempt the most likely criticisms.
   
   6. **Strengths** (what reviewers would praise):
      - This helps me know what to keep and what to emphasize in the cover letter
   
   Be brutal. A friendly review helps no one. I'd rather fix problems now than get a rejection letter in 4 months.

   Copy

   Tip: Every piece of feedback from a reviewer (real or simulated) is either a genuine improvement opportunity or a communication failure in your paper. If a reviewer misunderstands your methodology, the methodology section needs rewriting — even if you think it's clear. The reviewer's confusion IS the data point.

Recommended Tools for This Scenario

Claude

Freemium

Anthropic's AI assistant built for thoughtful analysis and safe, nuanced conversations

• 200K token context window for massive document processing
• Artifacts — interactive side-panel for code, docs, and visualizations
• Projects with persistent context and custom instructions

Get Started →

ChatGPT

Freemium

The AI assistant that started the generative AI revolution

• GPT-4o multimodal model with text, vision, and audio
• DALL-E 3 image generation
• Code Interpreter for data analysis and visualization

Get Started →

Elicit

Freemium

AI research assistant that automates literature reviews and data extraction

• Semantic search that understands research questions beyond keywords
• Automated data extraction from papers into structured tables
• Customizable extraction columns for any research domain

Get Started →

Consensus

Freemium

AI-powered academic search that synthesizes findings from peer-reviewed papers

• Natural language search across 200M+ academic papers
• AI-synthesized answers with inline citations
• Consensus Meter showing research agreement levels

Get Started →