The old advice to look for spelling mistakes and clunky phrasing simply doesn’t apply anymore. AI-generated emails often reference real, specific details about you, a conference you attended, a project you’re working on, a vendor you use, while the actual request feels slightly mismatched with the relationship.

A real colleague rarely opens with a callback to something specific before immediately pivoting to an urgent ask. That combination of accurate personal context plus an unusual request is one of the clearest tells left.

Trust that instinct if something reads correctly on the surface but feels just slightly off in how it’s framed. That gut sense is often more reliable than scanning for typos at this point.

One of the most consistent patterns across AI-generated scams is urgency delivered through a single channel. If a “CEO” or colleague emails an urgent request but never calls, texts, or follows up through your normal communication channels, that’s a strong signal something is wrong.

Real urgent requests inside most organizations tend to show up in more than one place. A wire transfer request that arrives only by email, with no Slack message or phone call backing it up, breaks that usual pattern.

Attackers rely on the pressure of urgency to short-circuit your instinct to double-check. Recognizing that pattern itself, urgency plus a single isolated channel, is often more useful than trying to spot a flaw in the writing.

“Quishing,” or QR code phishing, has become a favored workaround precisely because it sidesteps link-scanning defenses. Scammers increasingly embed QR codes in emails instead of clickable links, since many security scanners are built to analyze URLs, not images.

Once scanned, the code typically redirects to a fake login page designed to capture credentials, sometimes even a session cookie that can bypass multi-factor authentication entirely. The QR code itself feels more casual and less suspicious than a raw link, which is exactly the point.

This tactic shows up in fake invoice notices, subscription renewal warnings, and even calendar invites, areas where people don’t usually expect to encounter a QR code at all.

Adversary-in-the-middle, or AITM, attacks have evolved past simple fake login pages. These attacks can mirror legitimate login flows closely enough to capture both a password and an active session cookie, which can bypass multi-factor authentication after the fact.

This means entering your password correctly and even approving an MFA prompt no longer guarantees safety, if the page itself is sitting between you and the real service the whole time. The visual cues people were trained to look for, layout, logo, color scheme, are no longer reliable indicators on their own.

The domain in the address bar remains one of the few things attackers still struggle to perfectly fake, which makes it worth checking even when everything else looks correct.

Attackers are increasingly hiding malicious content inside platforms people already trust by default. Fake renewal notices have been found added directly into calendar invites, and scam pages have been hosted inside “invisible pages” within cloud documents to slip past standard web filters.

Because these live inside Google Calendar, shared documents, or other everyday productivity tools, they tend to bypass the scrutiny people normally apply to unfamiliar emails or links. The platform’s own reputation effectively vouches for content that has nothing to do with the platform itself.

This “reputation bypass” technique is becoming more common precisely because traditional spam filters are built to flag suspicious domains, not suspicious behavior happening inside an already-trusted one.