For years, the standard advice for spotting a scam message was to look for clumsy grammar, odd spelling and strange formatting. That advice has aged badly. AI writing tools now let scammers produce polished, convincing messages in seconds, tailored to the person reading them. The old tells are fading, but the underlying tricks haven't changed much — and those are what to watch for.

What AI has changed for scammers

Clean writing — messages read naturally, in the right tone, often in your own language

Personal details — information scraped from social media or past data leaks can be woven in to make a message feel targeted

Scale — one person can generate many varied messages, so you won't always see the same wording twice

Voice and video — cloned voices and altered video can impersonate someone you know or work with

The signs that still work

However well a message is written, it usually needs you to do something quickly, without checking.…