Oral Defense and Viva Preparation: Proving Authorship When Accused of AI Use

TL;DR

An oral defense (viva voce) is your most critical opportunity to prove authorship when accused of using AI. Success depends on process evidence—version histories, draft timestamps, research notes—not just verbal explanations. Institutions must prove misconduct by “balance of probabilities,” not absolute certainty. Start collecting evidence immediately: preserve Google Docs/Word version histories, document your research trail, and practice defending your arguments. Download our free evidence template at the end.

---

Introduction: The New Reality of AI Accusations

Universities worldwide are confronting a surge in AI-related academic misconduct cases. In the UK alone, AI accusation rates jumped from 1.6 to 5.1 cases per 1,000 students between 2022-23 and 2023-24—a 219% increase in one year (The Guardian investigation). This isn’t just about catching cheaters; it’s about false positives hitting innocent students, particularly non-native English speakers whose polished writing triggers detection algorithms.

When your university flags your work with an AI detector, the next step is rarely automatic punishment. Most institutions now require a viva voce—a live oral examination where you defend your work in person. This is your chance to demonstrate genuine understanding and present evidence of your writing process. But without proper preparation, even an innocent student can fail the viva due to nervousness, poor evidence presentation, or misunderstanding the process.

This guide combines current academic policies, legal frameworks, and proven defense strategies from universities worldwide. We’ll show you exactly what evidence matters, how to prepare, what questions to expect, and how to navigate special considerations for international students and disabilities.

1. Understanding the Viva Voce: What to Expect

What Is a Viva Voce?

“Viva voce” (Latin: “with living voice”) is an oral examination where you defend your work before a panel of examiners. While commonly associated with doctoral defenses, vivas are increasingly used for undergraduate and graduate work suspected of AI misuse. The format varies significantly by region:

Region	Typical Format	Purpose
UK	Private, examiner-driven, formal discussion	Both degree completion AND misconduct investigations
US	Public/semi-public, part of broader assessment	Primarily for doctoral completion; misconduct cases handled separately
Australia	Formal evidence-based interview	Standard procedure for academic misconduct cases
EU	Varies by country; often hybrid approach	Mixed usage for degrees and investigations

Source: University procedural documents from Imperial College London and City, University of London (2025).

The Viva Structure: Typical Agenda

Most vivas follow a similar pattern, whether for degree completion or misconduct investigation:

1. Opening statements (2-5 minutes): You summarize your work and its significance
2. Examiner questions (20-40 minutes): Focus on methodology, arguments, sources
3. Deep dive (10-20 minutes): Specific sections flagged for AI detection
4. Closing (2-3 minutes): You may have final statements

For AI accusation cases, expect intense scrutiny of:

• How you developed arguments
• Why you chose specific sources
• How you integrated different perspectives
• Your understanding of key terminology

How Long Does It Take?

Typical duration: 30-60 minutes. Some last 90 minutes for complex cases. You’ll usually have 15-30 minutes advance notice of specific questions (varies by institution).

2. Your Legal Rights and the Burden of Proof

The Institution Must Prove Misconduct

This is the most critical concept: you do not have to prove your innocence. Under UK and Australian law, universities must prove academic misconduct by “balance of probabilities” (more likely than not). In the US, the standard is usually “preponderance of evidence.” These are lower than criminal “beyond reasonable doubt,” but still require credible evidence.

What this means for you:

• An AI detector score alone is insufficient evidence
• The panel must show it’s more likely than not that you used AI improperly
• Your evidence doesn’t need to be 100% conclusive—just create reasonable doubt

Source: Imperial College London Academic Misconduct Policy (Section 4.2: Standards of Proof).

Your Rights During the Viva

You have specific procedural rights that vary by institution but generally include:

• Right to see the evidence: Request the specific AI detector results and which sections triggered flags
• Right to representation: Some institutions allow a student union advisor or legal counsel (especially for severe penalties)
• Right to present evidence: Bring version histories, notes, browser logs
• Right to appeal: If found guilty, you usually have 10-28 days to appeal to an independent committee
• Right to remain silent: You can decline to answer, but this may hurt your case

Check your university’s academic misconduct policy for exact details. The University AI Policies 2026 Global Tracker provides links to policies by country.

Common Procedural Violations to Watch For

Watch for these procedural errors that could invalidate the process:

• No specific evidence provided: If they won’t show you the detector report or flagged passages
• No opportunity to respond: Rushed process without your chance to present defense
• Unauthorized panel composition: Panel members without proper training in AI detection
• Failure to follow policy: Skipping required steps in your university’s published procedure

3. The Evidence Chain: What Actually Works

Your evidence is only as strong as its chain of custody. Panels want proof that shows a continuous writing process. Here’s what works, ranked by effectiveness:

Tier 1: Strongest Evidence

1. Version History from Google Docs/Microsoft Word

Digital version histories create an automatic timestamped record of every edit. This is your single most powerful evidence.

How to present it:

• Export version history as PDF (Google Docs: File → Version history → See version history → Export)
• Highlight key milestones: first draft, major revisions, final edits
• Show progression: sentence changes, paragraph additions, research integration

What it proves: You made incremental improvements over time, wrote in bursts of activity, and retained earlier versions with errors typical of human writing.

Note: Google Docs preserves versions for 30 days by default; Word keeps track for the document’s lifetime.

Source: University of Alberta AI Integrity Centre guidelines (2025).

2. Git/GitHub Commit History for Code or LaTeX

For technical students (CS, engineering, math), version control provides the most indisputable audit trail. Each commit shows:

• Timestamp
• Author (your account)
• Changes made (added/removed lines)
• Commit messages (reason for changes)

How to present:

• Take screenshots showing the commit timeline
• Include terminal output showing git log --oneline
• Highlight how you addressed feedback, fixed bugs, improved algorithms

This evidence is so strong that some universities recommend using Git for all academic work, not just programming.

Source: “Using Version Control (Git) as Evidence of Authorship in Academic Submissions” (academic integrity guidelines, 2025).

3. Draft Files with Timestamps

Even if you don’t use collaborative tools, having multiple draft files saved with file modification dates creates a timeline.

What to collect:

• Draft #1 (date)
• Draft #2 after supervisor feedback (date)
• Draft #3 after literature review additions (date)
• Final version (date)

How to enhance credibility:

• Include file properties showing creation/modification dates
• Keep a folder with all drafts in chronological order
• Note the storage location (OneDrive, Google Drive, USB) to show continuous access

Tier 2: Supporting Evidence

4. Research Notes and Browser History

Your research process—finding sources, taking notes, exploring concepts—shows you engaged deeply with the material.

What to present:

• Screenshots of browser tabs showing research sessions (with timestamps)
• Notes from journal articles (handwritten or digital)
• Citation management tools (Zotero, Mendeley) libraries showing when sources were added
• Annotated PDFs with highlights and margin notes

Pro tip: Export your browser history as CSV (Chrome: History → Export) to show research timeline.

5. Communication with Supervisors or Peers

Emails, chat logs, or meeting notes showing you discussed the work.

What to preserve:

• Emails asking for clarification on concepts
• Meeting notes with action items
• Feedback received and how you implemented it
• Questions you asked about methodology

Caution: Remove any irrelevant personal communications; only include academic discussions.

6. AI Disclosure Statements (If You Used AI Ethically)

If you used AI as a permitted tool (e.g., grammar checking, brainstorming), bring:

• Screenshots of prompts and outputs
• Your edits showing how you transformed AI suggestions
• Course syllabus or professor permission allowing AI use

Full disclosure of limited AI use can actually strengthen your credibility if done transparently.

Source: “Fair Use in Academia: How to Legally Use AI-Generated Content in Research Papers” (Paper-Checker Hub, 2026).

Tier 3: Contextual Evidence

7. Handwritten Notes or Outlines

Physical notes can demonstrate your thought process, especially for brainstorming sessions.

What’s effective:

• Rough sketches of argument structures
• Margin notes in textbooks
• Whiteboard or paper notes with timestamps (photos with date stamps)

8. Plagiarism/AI Self-Check Reports

If you ran your work through a checker before submission (like Paper-Checker or a Turnitin draft), bring those reports showing low AI percentages. This shows good faith effort.

What NOT to Rely On

• AI detector scores alone: These are screening tools with known false positives (especially for non-native speakers) Read more about false positives
• Memory-based testimony: Saying “I wrote it myself” without evidence carries little weight
• Appeals to emotion: Hard work and stress don’t prove authorship

4. Anticipating Questions: Practice for the Viva

The panel will ask questions designed to probe your depth of understanding. These fall into several categories:

Category A: Knowledge Demonstration

These test if you actually understand your own work.

Sample questions:

• “Explain the methodology you used in Chapter 3 in your own words.”
• “What are the limitations of your approach?”
• “How does your argument relate to [specific theory] we discussed in class?”
• “Define these terms in your own context: [key terms from your paper]”

How to prepare:

• Write one-paragraph explanations of each major section
• Practice explaining complex concepts to a non-expert (friend, family member)
• Identify 3-5 weaknesses in your own work and prepare honest responses

Category B: Process Evidence

These ask about your writing and research journey.

Sample questions:

• “Show us how you developed this argument over time.”
• “What was your research process for finding these sources?”
• “Walk us through your drafts—what changed between versions?”
• “How did you decide to include/exclude certain sources?”

How to prepare:

• Create a timeline of your work (research → drafting → revisions → final)
• Practice presenting your version history clearly
• Note specific challenges and how you overcame them

Category C: Source Integration

These test whether you truly engaged with your sources.

Sample questions:

• “How does Source X’s argument differ from Source Y’s?”
• “Why did you choose to cite Source Z for this point?”
• “What would Source A say about your conclusion?”
• “Explain this complex passage from Source B in your own words.”

How to prepare:

• For each major source: write a 2-sentence summary and note how you used it
• Identify debates between sources and your position
• Know your bibliography intimately

Category D: AI-Specific Probes

If the AI detector flagged specific sections, expect targeted questions:

Sample questions:

• “This paragraph has an unusually high perplexity score. Can you explain your thought process here?”
• “The detector suggests this section was machine-generated. How did you arrive at these exact phrases?”
• “Show us your notes for this specific argument.”

How to prepare:

• Anticipate which sections might be flagged (long complex sentences, sophisticated vocabulary)
• For each flagged section, prepare: your notes → drafts → understanding of why you phrased it that way
• Know the detector’s limitations explained here

Practice Drill: The 30-Day Viva Prep Plan

Weeks 1-2: Evidence Collection

• Gather all drafts, notes, browser history
• Export version histories and organize chronologically
• Create your evidence inventory (use our template below)

Weeks 3-4: Knowledge Review

• Write summaries of each chapter/section
• Identify 5 key arguments and supporting evidence
• Prepare responses to likely questions (top 20)

Week 5: Mock Vivas

• Practice with a friend acting as examiner
• Record yourself and critique delivery
• Time your responses (aim for 1-2 minutes per question)

Week 6: Final Preparation

• Print evidence documents (backup copies)
• Plan your outfit and logistics
• Prepare opening and closing statements

5. Special Considerations

International Students: Language Barriers

International students face disproportionate AI detection flagging because:

• Non-native writing patterns can appear “perplexed” to detectors (higher perplexity scores)
• Sophisticated vocabulary learned as a second language triggers false positives
• Cultural differences in academic expression affect authenticity

The statistics are clear: 68% of international students report language barriers affecting their viva performance, yet viva remains the fairest verification method because live interaction dispels detector myths.

Source: International student support offices, UK Universities (2025 survey).

Strategies for international students:

• Bring evidence of your language learning journey (ESL course completion, TOEFL/IELTS scores)
• Document glossaries of discipline-specific terms you learned
• Practice speaking about your work—even if nervous, comprehension matters more than accent
• Request language accommodations if needed: slower pace, clarification of complex questions

Important: Some universities have specific procedures for international students. Check your institution’s policy.

Students with Disabilities

If you have a disability affecting communication (speech disorders, anxiety, processing disorders), you have rights to accommodations.

Common accommodations:

• Extra time (often 50-100% more)
• Breaks during the viva
• Alternative formats for presenting evidence (written instead of oral responses)
• Support person or interpreter present

How to request:

• Contact your university’s disability services immediately upon receiving the accusation
• Provide medical documentation
• Submit formal accommodation request at least 5-7 days before the viva

Source: Disability Discrimination Act (UK), ADA (US), and equivalent EU directives.

6. Templates and Checklists

Evidence Collection Template

Use this spreadsheet to organize your evidence:

Evidence Type	File/Location	Date Range	Key Points	Presented?
Google Docs version history	Drive folder: /Thesis/Drafts	Sept 2024 – Feb 2025	Shows 50+ incremental edits; early drafts had weak arguments that improved	✅
Git commits (code chapters)	GitHub repo: username/thesis	Oct 2024 – Jan 2025	87 commits; avg 3 per day during writing sprints	✅
Research notes	OneNote: Literature Review	Nov 2024	Annotations on 45 sources with personal insights	✅
Supervisor emails	Gmail: label “thesis”	Dec 2024 – Feb 2025	12 exchanges discussing methodology	✅
Plagiarism check (pre-submit)	Paper-Checker report	Jan 30, 2025	Similarity 12%, AI 3% (well below threshold)	✅
Browser history export	Chrome: research.csv	Oct – Dec 2024	Shows 200+ research sessions on academic databases	✅

Viva Day Checklist

24 hours before:

• [ ] Print all evidence documents (3 copies: you, panel, backup)
• [ ] Charge laptop/tablet (optional, but useful for showing digital evidence)
• [ ] Plan route and arrival time (aim to be 15 minutes early)
• [ ] Review opening statement (2-3 minutes)
• [ ] Sleep well (no cramming)

Day of:

• [ ] Dress professionally (business casual)
• [ ] Bring: evidence binder, water bottle, notepad, pens
• [ ] Turn off phone (or set to silent/airplane)
• [ ] Arrive 15 minutes early
• [ ] Introduce yourself to panel politely
• [ ] Speak clearly and slowly (nervousness is normal)

During the viva:

• [ ] Listen fully to each question before answering
• [ ] Ask for clarification if unsure (better than guessing)
• [ ] Refer to evidence when appropriate: “If I may show you page 3…”
• [ ] Stay calm if challenged; respond with evidence, not emotion
• [ ] Take notes on questions you want to revisit

After the viva:

• [ ] Thank the panel
• [ ] Wait for decision (sometimes immediate, sometimes days-weeks)
• [ ] If found not guilty: request written clearance
• [ ] If found guilty: consult student union immediately for appeal

7. What Happens After the Viva?

Possible Outcomes

1. No misconduct found: Case dismissed. Your work stands as is. Request written confirmation for your records.
2. Minor academic malpractice: Reduced grade, required rewriting, or educational module completion.
3. Serious academic misconduct: Fail grade, course failure, or in extreme cases, expulsion. Appeal rights apply.

The Appeal Process

If you disagree with the decision:

1. Check deadline (usually 10-28 days)
2. Grounds for appeal: Procedural irregularity, new evidence, or unreasonable decision
3. Submit written appeal with supporting documentation
4. Attend appeal hearing (if required)
5. Await final decision (binding)

Contact your student union or academic advocacy office for free advice and representation during appeals.

Related Guides

If you’re dealing with AI accusations, these resources will help:

• False Positive AI Detection: Statistics, Causes, and Student Defense Strategies 2026 — Understand why detectors flag innocent work
• How to Appeal AI Detection False Positives: Complete 2026 Student Guide — Step-by-step appeal process
• University AI Policies 2026: Global Tracker for Students — Find your institution’s specific rules
• AI Use Policies by Country: 2026 Global Comparison for Students — How different countries handle AI accusations
• AI Detectors Explained: How Machine Learning Flags AI Writing — Technical deep dive into detector limitations

Summary: Your Next Steps

Facing an AI accusation is stressful, but systematic preparation dramatically improves your odds. Here’s your condensed action plan:

Immediately (within 48 hours):

1. Preserve all digital evidence (version histories, drafts, notes) immediately before files are lost or altered
2. Review your university’s academic misconduct policy online
3. Contact student union for advice and possible representation
4. Download our evidence template (below) and start filling it

Week 1:

1. Organize evidence chronologically
2. Write summaries of each section of your work
3. Identify 3-5 key arguments and your sources
4. Practice explaining your methodology without notes

Week 2:

1. Complete mock viva with a friend acting as examiner
2. Refine your evidence presentation (screenshots, printouts)
3. Anticipate tough questions and prepare responses
4. Print all documents and prepare binder

Day of Viva:

1. Arrive early, dressed professionally
2. Present evidence confidently
3. Answer questions with clarity and specificity
4. Refer to your evidence throughout

Remember: The burden of proof is on your institution. Your job is to create reasonable doubt by demonstrating a genuine, documented writing process.

Download Your Free Evidence Template

Preparing evidence is time-consuming. We’ve created a ready-to-use template to help you organize your version histories, drafts, research notes, and other process evidence in a format that satisfies university committees.

📋 Download Evidence Template (Excel)

We respect your privacy. No spam. Unsubscribe anytime.

Need Personalized Help?

Oral defenses are high-stakes. If you want expert guidance from academic integrity specialists, we offer:

• One-hour viva coaching: Mock viva with feedback on evidence presentation and responses
• Evidence review: Our team reviews your evidence pack before your panel hearing
• Policy interpretation: We help you understand your university’s specific procedures and your rights

🎓 Book a Viva Coaching Session

Limited slots available. Students who received coaching reported 89% success rate in 2025.

Last updated: March 5, 2026. All procedures and policies current as of 2026 academic year.