Research with AI | American University of Sharjah
Learning Pathway: Research

Research with AI

Accelerate every stage of your research: from finding literature to submitting your final paper, using powerful AI tools ethically and effectively.

Research-focused Full workflow 3-4 hrs total All levels welcome
Target Audience

Who is this for?

Whether you are writing your first essay or submitting to a peer-reviewed journal, this pathway is designed for you.

Undergraduate students

New to academic research? Learn to use AI as a reliable research companion: finding credible sources, understanding complex papers, and structuring your assignments with confidence.

  • Find and filter academic sources faster
  • Understand difficult papers in plain language
  • Brainstorm and outline your literature review
  • Format citations correctly, every time

Postgraduate students

Accelerate your thesis or dissertation. Use AI to manage large literature corpora, identify research gaps, and maintain momentum through every chapter of your work.

  • Map and synthesize large bodies of literature
  • Identify gaps and contradictions in the field
  • Analyze qualitative and quantitative data
  • Respond professionally to reviewer feedback

Faculty & researchers

From grant proposals to journal submissions, discover how AI can streamline your research pipeline and enhance collaboration with students.

  • Conduct rapid, comprehensive literature reviews
  • Accelerate data cleaning and pattern analysis
  • Draft and refine grant applications efficiently
  • Stay current with AI disclosure requirements
Interactive Process

The research journey

Select a phase below to explore specific tools, copy-paste prompts, and best practices for that stage of your research.

Phase 1 of 5

Define & discover

Craft a focused research question and use AI-powered discovery tools to find the most relevant, peer-reviewed literature in your field: faster than ever before.

What you will learn

  • Narrow a broad topic into a specific question
  • Generate effective search keywords using AI
  • Discover papers beyond standard database searches

Key concepts

  • Research question vs. research hypothesis
  • Citation mapping and co-citation analysis
  • Identifying research gaps in existing literature
For Undergrads: Start by exploring Consensus to find peer-reviewed papers. Use the AI summaries to quickly assess relevance.

Beginners start here: identifying your focus

If you have never researched before, the hardest part is turning a broad idea into an answerable question. Follow this proven sequence:

  • Step 1: Start Broad You are interested in artificial intelligence. That is too broad to research.
  • Step 2: Ask an AI Chatbot for Challenges Ask a tool like ChatGPT: What are the current major challenges in deploying AI in public healthcare?
  • Step 3: Identify the Research Gap The AI might tell you that patient data privacy is a significant hurdle preventing hospitals from adopting AI diagnostics.
  • Step 4: Refine the Question Your new, focused question becomes: How do federated learning models mitigate patient data privacy risks in hospital diagnostic AI systems?
  • Step 5: Verify with Academic Tools Now, take that specific question to Consensus or Elicit to find real, peer-reviewed papers that explore this exact topic.

Tools for phase 1

Consensus Freemium

An AI-powered academic search engine trained specifically on peer-reviewed research. Ask a research question in plain English and receive a ranked list of relevant papers with AI-generated evidence summaries.

Visit Consensus
Elicit Freemium

Searches over 125 million academic papers and extracts structured information such as population, intervention, outcomes, and study design into a clean table. Perfect for comparing methodologies.

Visit Elicit
Research Rabbit Free

A visual literature discovery tool. Start with one or more seed papers, and it visualizes citation networks, showing you how papers relate to each other to find seminal works.

Visit Research Rabbit

Copy-paste prompts

Research Question Prompt
ChatGPT / Claude
I am working on a research topic about [YOUR BROAD TOPIC] for a [undergraduate essay / postgraduate thesis / journal article] in the field of [YOUR DISCIPLINE]. Help me narrow this into 3 specific, researchable questions. Explain why each is feasible and what methodology suits it best.
You can use the copy paste prompt above on any LLM such as Gemini, Claude, Grok, etc.
Hands-On Exercise
Running Scenario: Used Across All 5 Phases

Research topic: social media use & university student academic performance

Throughout all five phases you will work on the same real topic. By Phase 5 you will have gone from a vague idea all the way to a submission-ready bibliography. Follow along step by step: each exercise builds on the last.

Try it now: discover papers with Consensus & Elicit

Follow the steps below to find and map relevant literature.

~20 mins
  1. 1
    Search Consensus for peer-reviewed papers. Enter this specific query to see how Consensus differs from a standard search engine.
    Query to copy Does social media use negatively affect academic performance in university students?
    Launch Consensus
  2. 2
    Extract structured data using Elicit. Run the same query in Elicit. Notice the table format. Click "Add column" and select "Sample size" to extract this data point automatically. Launch Elicit
  3. 3
    Map citations with Research Rabbit. Take the most relevant paper from your search and paste its title into Research Rabbit to find connected seminal works. Launch Research Rabbit
Completion Checklist
  • Reviewed 5 Consensus summaries
  • Extracted sample sizes in Elicit
  • Mapped 1 paper in Research Rabbit
  • Drafted specific research question
0 / 4 complete
Expected Output
  • List of 5 peer-reviewed papers
  • 1 citation network map
  • Refined research question ready for Phase 2
Jump to Phase:
Phase 2 of 5

Read & synthesize

Use AI to digest dense academic papers quickly, extract key arguments, compare multiple studies, and build a coherent understanding of your literature landscape.

What you will learn

  • How to upload papers and query them with AI
  • Extracting methodology, findings, and limitations
  • Building a thematic synthesis from diverse sources

Critical reminders

  • Always verify AI summaries against the original
  • Do not cite a paper you have not personally reviewed
  • Use AI to understand, not to replace, your reading

Beginners start here: reading dense papers

Academic papers can be intimidating. Here is how to use AI to break them down without getting overwhelmed:

  • Step 1: Read the Bookends First Always read the abstract and the conclusion yourself before uploading the paper to any AI tool. This grounds your understanding.
  • Step 2: Upload and Extract Methodology Upload the PDF to Claude or NotebookLM and ask: "Explain the methodology of this paper to an undergraduate student. Skip the complex math equations and focus on how the experiment was actually run."
  • Step 3: Query the Main Findings Ask the AI: "What are the 3 most significant findings of this study, and what are the limitations the authors themselves admit?"
  • Step 4: Verify Against the Text If the AI claims a specific finding, return to the original PDF and verify the claim in the Results section before writing it down in your notes.

Tools for phase 2

NotebookLM Free

Google's AI research assistant allows you to upload up to 50 documents and query them collectively. It generates cross-document summaries and provides source-verified footnotes.

Visit NotebookLM
Claude Freemium

Anthropic's LLM is currently superior at parsing long, complex academic texts. Upload a challenging paper and ask it to break down the methodology in simpler terms.

Visit Claude

Copy-paste prompts

Cross-Paper Synthesis Prompt
NotebookLM
I have provided you with [NUMBER] research papers on [YOUR TOPIC]. Synthesize them and produce: 1. Common Themes 2. Key Agreements & Disagreements 3. Methodological Patterns 4. Remaining Questions
You can use the copy paste prompt above on any LLM such as Gemini, Claude, Grok, etc.
Hands-On Exercise
Continuing from Phase 1

You now have 5 papers. time to understand them without reading every word.

You will upload those papers to NotebookLM and query Claude individually on the most complex one. By the end you will have a structured synthesis table ready to inform your literature review.

Try it now: multi-paper synthesis

Process your downloaded PDFs using AI reading assistants.

~25 mins
  1. 1
    Download PDFs of your top 3 papers. Save the papers you identified in Phase 1 to your local computer.
  2. 2
    Upload papers to a new NotebookLM notebook. Create a notebook called "Social Media & Academic Performance" and add the PDFs as sources. Launch NotebookLM
  3. 3
    Run a synthesis query across all documents. Ask NotebookLM to extract findings, sample sizes, and limitations, then compare the papers. Click footnotes to verify.
  4. 4
    Deep dive into complex methodology with Claude. Take the hardest paper to Claude to break down its statistical methods in plain English. Launch Claude
Integrity Check: Use AI outputs as reading aids, not text to copy into your assignment. Write your own synthesis notes.
Completion Checklist
  • Uploaded 3 PDFs to NotebookLM
  • Ran synthesis query & verified footnotes
  • Queried Claude on hardest paper
  • Wrote personal synthesis notes
0 / 4 complete
Expected Output
  • 3-paper comparison table
  • Verified source claims
  • Plain-English statistical notes
Jump to Phase:
Phase 3 of 5

Collect & analyze data

Design stronger data collection instruments and use AI to identify patterns, interpret results, and generate visualizations from your quantitative or qualitative data.

What you will learn

  • Use AI to design and critique survey questions
  • Identifying patterns and generating visualizations
  • AI-assisted thematic coding for qualitative data

Data privacy rules

  • Never upload PII to public AI tools
  • Anonymize all data before sharing with AI
  • AI supplements, does not replace, statistical expertise

Beginners start here: preparing your data

Before asking an AI to analyze your dataset, you must ensure it is clean and safe to share:

  • Step 1: Strip Identifying Information Remove all names, email addresses, phone numbers, and physical addresses from your spreadsheet. Use participant IDs (e.g., Participant_01) instead.
  • Step 2: Simplify Column Headers Ensure your top row contains clear, concise variable names (e.g., Age, Reaction_Time_ms) without special characters.
  • Step 3: Handle Missing Data Decide how you will mark blank cells. Do not leave them completely empty; use a consistent marker like NA or 999.
  • Step 4: Prompt the AI Analyst Upload the clean CSV to a tool like Julius AI and start with a simple prompt: "Describe this dataset. What is the distribution of the main variables, and are there any obvious outliers?"

Tools for phase 3

Julius AI Freemium

A purpose-built AI data analyst. You can upload CSV, Excel, SPSS, or Stata files and ask it to run analyses in plain English. It can generate publication-quality charts.

Visit Julius AI
ChatGPT (Data) Freemium

The paid tier of ChatGPT includes a Python environment. When you upload a dataset, it writes and executes Python code to clean your data and run statistical tests.

Visit ChatGPT

Copy-paste prompts

Dataset Exploration
Julius / ChatGPT
I have uploaded a dataset related to [YOUR TOPIC] containing [BRIEF DESC]. Please provide descriptive statistics, identify missing data/outliers, show distribution of [KEY VARIABLE], and suggest appropriate statistical tests.
You can use the copy paste prompt above on any LLM such as Gemini, Claude, Grok, etc.
Hands-On Exercise
Continuing from Phase 2

Imagine you surveyed 50 students. let's build and analyze that dataset right now.

You will create a small practice CSV file, upload it to Julius AI, and ask it to find patterns. You will see exactly how a real data analysis conversation with AI works: before you ever collect real data.

Try it now: analyze sample data

Use Julius AI to analyze a provided mock dataset.

~30 mins
  1. 1
    Create practice dataset. Copy the data below and save as social_media_study.csv.
    csv 
    ParticipantID,Age,Gender,SocialMedia_HrsPerDay,GPA,StudyHrs_PerWeek,Platform
P001,19,F,4.5,3.1,18,TikTok
P002,21,M,1.0,3.8,26,Instagram
P003,20,F,6.0,2.7,12,TikTok
P004,22,M,2.5,3.5,22,YouTube
P005,19,F,5.5,2.9,14,TikTok
P006,23,M,0.5,3.9,30,None
P007,20,F,3.0,3.3,20,Instagram
P008,21,M,7.0,2.4,10,TikTok
P009,19,F,2.0,3.6,24,YouTube
P010,22,M,4.0,3.0,16,Instagram
  2. 2
    Upload to Julius AI. Drag and drop your saved CSV file into a new Julius AI chat. Launch Julius AI
  3. 3
    Request descriptive stats & visualizations. Ask Julius to create a scatter plot of Social Media Hours vs. GPA and describe the relationship. Then, request a correlation analysis and a group comparison chart.
Privacy Check: Only practice data used here. Never upload real participant PII to public AI tools.
Completion Checklist
  • Created CSV file
  • Uploaded to Julius AI
  • Generated scatter plot
  • Ran correlation analysis
  • Noted sample size limitations
0 / 5 complete
Expected Output
  • Scatter plot & bar chart
  • r-value and p-value explanation
  • Understanding of limitations
Jump to Phase:
Phase 4 of 5

Write & draft

Use AI to create outlines, draft every section of your paper, refine your academic voice, and overcome writer's block while keeping your intellectual contribution front and centre.

What you will learn

  • Create strong, logical paper outlines
  • Improve academic tone, clarity, and concision
  • Maintain your own voice while using AI assistance

Academic integrity rules

  • Never submit AI-generated text as your own
  • Always revise and own every sentence you submit
  • Check your assignment brief before using AI

Beginners start here: structuring your draft

Academic writing is formulaic. Most research papers follow the IMRaD structure. Use AI to outline these sections before you type a single paragraph:

  • Step 1: Introduction Outline Ask AI: "Help me outline the introduction for my paper on [Topic]. It should cover the broader context, the specific problem, and end with my thesis statement."
  • Step 2: Methods Description The Methods section is just a recipe. Write down step-by-step what you did, and use AI to elevate the tone: "Rewrite these bullet points into a formal academic methods section."
  • Step 3: Results Summary This section only contains facts, no opinions. Paste your statistical output and ask AI to help you describe the data trends objectively.
  • Step 4: Discussion Structuring This is where your voice matters most. Use AI as a sounding board: "Here are my findings. Do they align with the current literature, and what are the implications?"

Tools for phase 4

Jenni AI Freemium

An AI writing assistant specifically designed for academics. As you write, it offers in-text autocomplete suggestions. It also has a built-in citation feature.

Visit Jenni AI
Claude Freemium

Because of its nuanced understanding of tone, Claude is excellent for the editing phase. You can paste a rough draft and ask it to improve the academic register.

Visit Claude

Copy-paste prompts

Academic Tone Improvement
Claude / ChatGPT
Improve the academic tone of this passage from my paper in [YOUR DISCIPLINE]: [PASTE PARAGRAPH] Improve formal register, eliminate informal phrasing, tighten wordy sentences, ensure appropriate hedging, and preserve my original argument.
You can use the copy paste prompt above on any LLM such as Gemini, Claude, Grok, etc.
Hands-On Exercise
Continuing from Phase 3

You have your literature and your data. now write the first real draft of your introduction.

You will use Claude to go from rough bullet points to a polished academic introduction, and then practise improving a weak paragraph using the tone improvement prompt.

Try it now: draft and refine

Generate outlines and improve writing tone using AI.

~35 mins
  1. 1
    Generate an outline with Claude. Provide your research question and key findings to Claude and request a detailed IMRaD outline. Launch Claude
  2. 2
    Draft text manually. Write out a rough draft of your introduction based on the outline. Do this yourself before seeking AI help.
  3. 3
    Refine tone and flow. Paste your draft into Claude using the "Tone Improvement" prompt. Review the changes carefully.

    Before

    Lots of students use TikTok and it's bad for grades.

    After AI Refinement

    Emerging evidence suggests frequent TikTok use among undergraduates correlates with reduced academic performance.

  4. 4
    Apply the read-aloud test. Read the AI-refined text aloud. If it sounds overly robotic or unnatural, edit it back to match your own voice.
Integrity Check: Do not copy-paste AI text wholesale. You must understand and own every argument in your paper.
Completion Checklist
  • Generated outline
  • Manually drafted introduction
  • Refined tone with Claude
  • Performed read-aloud test
  • Drafted Results paragraph
0 / 5 complete
Expected Output
  • Detailed paper outline
  • Refined introduction draft
  • Clear distinction of personal vs AI edits
Jump to Phase:
Phase 5 of 5

Cite & finalize

Manage your references with precision, format citations correctly, proofread your final draft, and prepare a polished submission, from cover letter to response-to-reviewers.

What you will learn

  • Setting up Zotero or Mendeley for citations
  • Using AI to proofread for grammar and flow
  • Writing a professional response to peer reviewers

Citation integrity

  • Always verify AI-generated citations
  • Never cite a paper you have not verified exists
  • Zotero and Mendeley are far more reliable than AI

Beginners start here: pre-submission checklist

Before you submit your final paper, complete this rigorous review process to avoid academic integrity issues:

  • Step 1: Verify Every Citation Never trust an AI to generate a citation from scratch. Verify that every single reference in your bibliography actually exists in a database like Google Scholar.
  • Step 2: Automate Formatting Do not format references by hand. Use a tool like Zotero or Mendeley to automatically generate your bibliography in perfect APA, MLA, or Chicago style.
  • Step 3: Final Flow Check Paste your final draft into Claude or ChatGPT and ask: "Proofread this for grammatical errors and awkward transitions only. Do not rewrite my arguments."
  • Step 4: The Read-Aloud Test AI writing can sometimes sound overly formal or robotic. Read your paper out loud. If a sentence sounds like something a human would never say, rewrite it in your own words.

Tools for phase 5

Zotero Free

The gold standard open-source reference manager in academia. It acts as your personal library. The browser extension captures paper metadata with one click.

Visit Zotero
Mendeley Free

A reference manager built by Elsevier. It offers robust PDF organization, built-in annotation tools, and features for sharing libraries with collaborators.

Visit Mendeley

Copy-paste prompts

Reviewer Response
Claude / ChatGPT
Write a professional response to this reviewer comment: COMMENT: [PASTE COMMENT] MY RESPONSE: [WHAT YOU DID] Draft a formal response thanking them, stating agreement/disagreement, describing specific changes, and justifying the choice with evidence.
You can use the copy paste prompt above on any LLM such as Gemini, Claude, Grok, etc.
Hands-On Exercise
Final Phase: Completing the Journey

You have drafted your paper. Now build a perfect bibliography and do a final AI-assisted proofread.

In this final exercise you will install Zotero's browser connector, save your Phase 1 papers into a library with one click, generate a flawless APA bibliography, and run a final integrity check.

Try it now: finalize & cite

Automate your bibliography and perform a final proofread.

~30 mins
  1. 1
    Setup Zotero. Install Zotero and its browser connector to quickly save metadata. Get Zotero
  2. 2
    Save references and build bibliography. Save your Phase 1 papers. Select them in Zotero, right-click, and "Create Bibliography" in APA format.
  3. 3
    Verify citations manually. Cross-reference the generated citations against original sources to ensure accuracy.
  4. 4
    Final AI Proofread. Use Claude for a final grammar and flow check using the strict proofreading prompt to prevent unapproved edits.
  5. 5
    Draft AI Use Statement. Include a disclosure detailing how AI was utilized throughout your research process.
Integrity Check: Always verify citations. A fabricated citation is a severe academic offense.
Completion Checklist
  • Installed Zotero
  • Generated automated bibliography
  • Verified all citations
  • Completed final proofread
  • Included AI Use Statement
0 / 5 complete
Expected Output
  • Properly formatted bibliography
  • Verified sources
  • Error-free final document
Jump to Phase:
Guidelines

Academic integrity & responsible use

Using AI in research comes with important responsibilities. Understanding these principles protects your academic work, your integrity, and the quality of your research.

Transparency

Always disclose when and how you have used AI in your research. Most journals and institutions now require explicit AI use statements in submitted work.

Verification

AI models can generate plausible-sounding but entirely incorrect information. In research, unchecked AI outputs can undermine the credibility of your entire work.

Intellectual ownership

Your research contribution is your own thinking: the argument, the interpretation, the insight. AI can assist your process, but the intellectual core must be yours.

Data privacy

Be extremely careful about what data you share with AI tools. Unpublished research data, participant information, and proprietary datasets all carry privacy obligations.

AI hallucinations in research are a serious risk

AI models frequently generate references that appear legitimate but do not exist, with realistic author names, journal names, volume numbers, and DOIs. Submitting a paper with hallucinated citations is a significant academic integrity violation. Always verify every reference in a database like PubMed, Scopus, or Google Scholar before including it in your work.

AUS AI policy for research

The American University of Sharjah has guidelines governing the use of AI tools in academic work. These guidelines cover disclosure requirements, approved tools, and discipline-specific rules.

Read the AUS AI Policy
Support

Frequently asked questions

Common questions regarding the use of artificial intelligence in research workflows.

Can I use AI tools for my research assignments at AUS?
It depends on the specific assignment and your instructor's guidelines. Always check your course syllabus first. The AUS AI Policy provides university-wide guidelines, but individual instructors may set stricter rules. When in doubt, ask your instructor before using any AI tool.
How do I properly cite AI tools in my research paper?
Most major citation style bodies (APA, Chicago, MLA) now have guidance on citing AI tools. Additionally, many journals and institutions now require an "AI Use Statement" in your methods section describing which tools were used, for which parts of the research process, and in what capacity. A sample AI Use Statement is included in the Phase 5 hands-on exercise above.
Is it safe to upload my unpublished research data to AI tools?
Exercise extreme caution with unpublished research data. Never upload personally identifiable participant data to any public AI tool. Always anonymize data thoroughly before any AI-assisted analysis. The Phase 3 hands-on exercise uses a practice dataset specifically to let you learn the analysis workflow before you ever touch real participant data.
What if the papers I need are behind a paywall?
Before paying for a paper, try these free routes in order: (1) check your AUS library database access; most major journals are covered, (2) search the paper's title on Google Scholar and look for a "PDF" link from a university repository, (3) use the Unpaywall browser extension, which finds legal open-access versions automatically, (4) email the corresponding author directly; researchers are almost always happy to share their work. Never use illegal piracy sites for academic papers.
How do I know if an AI summary of a paper is accurate?
The most reliable AI tools for this are those that show source citations, specifically NotebookLM and Elicit. Both link every claim back to the exact sentence in the original document. After receiving any AI summary, click each footnote and read the original passage yourself. If a tool does not show citations (like a general chatbot), treat its summaries as a rough orientation only; never as a citable source of information about the paper.