For years the advice was: look for typos. That advice is now dangerous. Scammers use the same AI writing tools everyone else does, so a fake bank email reads exactly like a real one, in any language, and can mention your name, your town or your recent purchase.
Live from Inbox Hero
Real totals across every inbox Inbox Hero protects: emails checked, flagged as suspicious, scams blocked and AI-written emails spotted. Updated Oct 28, 9:01 AM PT.
| Checked | Flagged | Blocked | AI-written | |
|---|---|---|---|---|
| Today | 1,271 | 4 | 3 | 484 |
| Last 30 days | 36,743 | 26 | 11 | 1,514 |
What AI gives scammers
- Perfect writing. No spelling or grammar mistakes, in the right tone for a bank, a lawyer or a grandchild.
- Personal detail at scale. Details from data breaches and social media are woven in automatically, for thousands of people at once.
- Voice and video. A few seconds of someone's voice from a video is enough to clone it for a follow-up call.
- Instant replies. If you answer, an AI keeps the conversation going convincingly.
What still gives an AI scam away
AI can fake the words. It cannot fake the plumbing. Every scam still has to send you somewhere or get you to do something, and that is where it shows:
- Where the link really goes, as opposed to what it says.
- The sending domain and whether it passes the real company's authentication (SPF, DKIM, DMARC).
- A reply-to address that differs from the sender.
- The ask: money by gift card, wire or crypto, a password, or a code.
This is how Inbox Hero works: it ignores how polished the writing is and checks the links, the domains, the attachments and the authentication, then has AI confirm only the cases that already look wrong. Inbox Hero also labels emails that look AI-written, so you know when a "personal" note was produced by a machine.
Protect the people most at risk
Older family members are the main target of AI scams because the old red flags were what they were taught to look for. Put a scam filter on a parent's inbox and agree on a family code word for any call that claims to be an emergency.
Frequently asked questions
What is AI phishing?
Phishing emails written or personalised by AI. They have no spelling mistakes, sound like the real sender, and can include personal details, which makes them much harder to spot by reading alone.
How do I spot an AI-written scam email?
Stop judging the writing. Check where links really go, whether the sender domain is exactly right, whether the reply-to address differs, and whether the email asks for money, a password or a code. Those do not change with AI.
Can AI clone a family member's voice?
Yes. A short clip from social media can be enough. Agree on a family code word and call back on a number you already have before sending money.
Can software detect AI scam emails?
Yes, by checking what AI cannot fake: links, domains, attachments and sender authentication. Inbox Hero does this on every email from a stranger and quarantines clear scams.
Take back your inbox today
Connect Gmail or any mailbox in one click. Inbox Hero files the noise, drafts your replies, and quarantines phishing. It is free for 30 days ($0 today).
Get started free