Skip to main content
Skip to content
Menu

AI Vulnerability Beginner 3 parts

Audit AI outputs for fabricated references before sharing work publicly

You'll have a personal verification protocol you run on any AI-generated research, citations, or sources before publishing or submitting work.

What you'll make

Citation Verification Protocol

  1. Red Flag Checklist: warning signs in AI citations
  2. Verification Steps: your testing workflow
  3. Quality Thresholds: when to accept, edit, or reject AI research

The 3 parts

Each part ends with a check you can see.

  1. Catch AI Citation Fabrications

    Test AI-generated citations against actual sources to discover how often AI fabricates references

    Done when

    You've caught at least 2 fabricated citations and verified at least 3 real ones, with evidence of what makes each fabricated citation look convincing

  2. Build Your Verification Workflow

    Design a step-by-step process for checking AI research that you can complete in under 10 minutes per source

    Done when

    You have a 4-6 step verification workflow that caught fabrications in testing and takes under 10 minutes per citation

  3. Set Your Quality Thresholds

    Define when AI research output meets your standards for public sharing versus when it needs more work

    Done when

    You have specific criteria for accepting, editing, or rejecting AI research output, tested against real examples

    Capstone: you put together your Citation Verification Protocol.

Where this helps

An elective: a single lesson that sits outside rtwo's job certificates and doesn't count toward one.

About this lesson

What you'll be able to do

  • Test every AI-generated citation before submission
  • Distinguish reliable from fabricated sources within 5 minutes
  • Build a verification protocol you can run on any AI research output
  • Defend work integrity when using AI as a research assistant

Who it's for

  • Researchers who use AI for literature reviews
  • Content creators who cite sources in their work
  • Anyone submitting AI-assisted work for publication or professional review

What it won't cover

  • How to write better prompts
  • AI detection tool development
  • Institutional policy recommendations
  • Citation format guidelines

Skills you'll practise

  • Citation verification
  • Source validation
  • AI output quality assessment
  • Systematic fact-checking