Scale, polish, imitation
Polish is now free
For years the advice was to look for bad spelling and odd grammar. AI writing tools removed that clue overnight. A scam email can now be fluent, polite, in your language and in the house style of your bank or delivery company. The same tools help millions of people write genuine messages every day, so polish proves nothing in either direction.
Scale and personalisation
One scam template can become thousands of individual versions, each mentioning your name, your town, a recent purchase or your employer, pulled from leaked data or public profiles. Replies can be handled by chatbots that keep the conversation going for days. More attempts reach more people, and each one feels more like it was written for you.
Imitation of people
A few seconds of someone's voice from a video or voicemail can be enough to clone it convincingly. Photos can be turned into short videos, and faces swapped on live video calls. That makes "it's me, I need help" calls and fake video meetings with a "boss" or "bank manager" more convincing than ever. Lessons 4 and 6 cover these in detail.
What did not change
Scams still need you to do something: pay, move money, share a code, log in through their link, or install something. That is the moment you control. Checking through a route you already trust (an app you installed, a number on your card, a number you already have saved) defeats AI-written messages and cloned voices alike, because the scammer cannot fake your own phone book.
Think about the last suspicious message you received. Would you have spotted it if it had no spelling mistakes?