Deepfake-spotting guide
You don’t need a lab to handle this. You need a pause, a call-back, and a plan for when a clip starts travelling. This is general guidance, not a detection product. Leave your email for a short PDF you can share with family or the team.
Step 1
Pause before you trust a face
If a video or a photo comes with a request for money, secrecy, or a login, treat the media as untrusted until you confirm it another way. A convincing face is not proof.
Step 2
Call back on a number you already have
Use a contact already in your phone, a meeting you scheduled, or a colleague in the room. Do not use a number or a link that arrived with the clip.
Step 3
Look, but don’t rely on looking
Odd lighting, missed glasses or jewellery, waxy skin, or lip-sync slips can still give a fake away. Better generators keep closing those gaps, so clues are a bonus, not a defence.
Step 4
Use a code word for family and for payments
Agree a private word. No urgent transfer, one-time code, or “keep this between us” request goes through without it and a call-back.
Step 5
If a clip is about your organisation
Say so on a channel you already own, and point people there. Don’t debate the fake only in the thread that is spreading it.
Get the PDF
Email plus a consent tick, that’s it. We never ask for a password.
Read the longer explainer on deepfakes or set up a family or team code word.