AI writing tools have made scam emails harder to spot the old way — poor grammar and awkward phrasing used to be reliable red flags, and increasingly they aren’t. Here’s what to check instead.
Why the old advice is less reliable
For years, one of the most common tips for spotting a scam email was to look for spelling mistakes and awkward, poorly translated phrasing. AI writing tools have largely closed that gap — a scam email can now read as fluently as a legitimate one, so grammar alone is no longer a dependable signal.
What still gives scams away
A few signals remain reliable regardless of how well-written the email is:
- The sender’s actual email address (not just the display name) rarely matches the organization it claims to be from
- Urgency and pressure to act immediately, especially around money, passwords, or account suspension
- Links that don’t match the real organization’s domain when you hover over them
- Requests to “verify” or re-enter a password or payment details via a link, rather than by logging in directly
Newer red flags to watch for
AI-generated scam emails sometimes introduce their own tells: an oddly generic tone that doesn’t quite match how the real organization normally writes, overly perfect but slightly impersonal phrasing, or small factual inconsistencies (wrong account details, a slightly wrong company name) that a careful human writer at the real company would be less likely to make.
When in doubt, verify independently
The most reliable check remains independent verification: rather than clicking a link or replying to the email, go directly to the organization’s official website (typed manually, not clicked) or call a phone number you already know is correct, and confirm the message through that separate channel.
Frequently asked questions
Can spam filters catch AI-written scam emails?
Modern spam filters look at more than just writing quality — sender reputation, links, and behavioral patterns still help catch many AI-written scams, though no filter is perfect.
Is good grammar in an email proof that it’s legitimate?
No. AI tools have made fluent, error-free writing available to scammers as well, so grammar alone is no longer a reliable indicator either way.
What’s the single most reliable way to check a suspicious email?
Verify independently: contact the organization through a channel you already trust (a phone number you looked up yourself, or their official website typed directly) rather than anything in the email itself.