Lion Cipher Lab
Spot deepfakes and manipulated media: a defensive verification workflow
- AI Security
- Beginner
- ~30 min
- Free · no account required
Why it matters
Convincing AI-generated and manipulated images, audio, and video are now cheap to produce and hard to catch by eye, and they are already being used for real harm: impersonating executives to authorize fraudulent payments, cloning voices for scam calls, and flooding public conversation with fabricated 'evidence' during disinformation campaigns. The instinct most people fall back on — staring harder to judge whether something looks fake — is exactly the instinct that fails, because the visual giveaways are unreliable and disappearing. What actually protects you is a verification habit: check…
What you'll learn
- media verification workflow
- reading content provenance (C2PA) as the durable signal
- reverse-image search and source tracing
- judging automated deepfake detectors honestly
- defensive framing of media fraud and disinformation
Steps in this lab
- The problem: this is fraud, not a novelty
- Why 'just look for the tells' fails (and why we will not teach it)
- The durable signal: provenance, not appearance
- Workflow step 1: check provenance first
- Workflow step 2: trace it to a source with reverse-image search
- Workflow step 3: verify the source and the context
- Where automated 'detectors' fit: one weak input among many
- The habit to carry out of this