DooDooLamb News
These AI Scams All Have Red Flags. Here's How to Spot Them
Brief published June 22, 2026 ยท Original source published June 20, 2026
Original reporting by Dianna Gunn at cnet.com.
Automated brief. Verify important details at the original source.
AI Has Made Scammers Dramatically Better at Their Jobs, Which Is the One Productivity Gain Nobody Asked For
The robots finally came for our jobs. Not the jobs anyone predicted, but the noble profession of Nigerian prince impersonation.
For decades, we laughed at phishing emails. The broken grammar, the implausible urgency, the prince who specifically needed your bank account to temporarily hold seventeen million dollars. That was the immune system working. The obvious scam was obvious on purpose, a filter to find only the most credulous targets. AI just disabled that filter for everyone.
The CNET piece making the rounds this week is a useful, earnest guide to spotting AI-assisted scams, and reading it carefully produces a specific kind of dread that no horror movie has managed in years. The core problem is not complicated: the large language models that your favorite startup is using to "10x your productivity" are the exact same technology that scammers are now using to 10x their productivity. The irony is so complete it has started filing its own taxes.
The scams described are not exotic. They are the same old cons, inheritance schemes, fake book clubs, suspicious prize notifications, romance grifts, but now dressed in prose that would pass a high school English class. This is what happens when you democratize access to a very good writing assistant and discover, to nobody's real surprise, that criminals are also early adopters. The cybersecurity industry spent years telling us to look for typos as a red flag. AI ate that advice and asked if there was anything else it could help with today.
What makes this genuinely insidious is the personalization angle. Modern scam operations are not just using AI to write cleaner sentences; they are using it to scrape publicly available information about you and weave it into the pitch. Your LinkedIn says you're a logistics manager in Phoenix? Wonderful, here is an extremely plausible email about a shipping irregularity requiring your immediate attention and, incidentally, your Social Security number. This is what the industry calls a "spear phishing" attack, which is distinguished from regular phishing the way a sniper is distinguished from someone throwing rocks: same goal, considerably more precision, much harder to dodge. The AI layer means that crafting one of these targeted messages no longer requires a skilled social engineer spending hours on research. It requires a prompt and about forty-five seconds.
The red flags that survive in this new environment are behavioral rather than linguistic. The grammar is clean now. The logo looks right. The sender name is plausible. What remains detectable is the ask. Legitimate institutions, this is worth tattooing somewhere visible, do not cold-contact you to request sensitive information. Your bank does not email you asking you to confirm your password. The IRS does not text you. No book club on earth needs your full credit card number to send you a free copy of something. The urgency is artificial, the deadline is fake, and the offer exists only as a vector. Any communication that combines an unsolicited contact with a time-sensitive demand for personal data is not a coincidence; it is a recipe that scammers have been following since before the internet and have now simply had printed on nicer stationery.
The advice in pieces like this one always lands in the same place: verify independently, contact companies through their official channels, and treat any unsolicited request for sensitive information with the suspicion you would reserve for a stranger handing you a USB drive in a parking lot. This is correct advice. It is also advice that requires a level of sustained paranoia that is genuinely exhausting to maintain, which is, of course, exactly what the scammers are counting on. They do not need to fool you every time. They need to catch you on a Tuesday afternoon when you are tired and three meetings deep and the email looks plausible enough that your guard drops for thirty seconds.
What This Actually Means for People Who Use the Internet
The uncomfortable reality check here is that the defense is not keeping pace with the offense. Email providers are improving their filters, and yes, Gmail catches a lot. But the CNET author's observation that scam emails are increasingly making it through is the data point that matters. Every time a defensive system improves, the training data for circumventing it also improves, because the scammers who get through are providing exactly the examples needed to get through again. This is not a solvable problem in the way we usually mean "solvable." It is an arms race between automated deception and automated detection, and the humans in the middle are not getting faster.
For developers building anything that touches user communication, this is also a design problem you now own whether you asked for it or not. Every email notification, every account alert, every customer service message you send is now operating in an environment where users are (rightly) suspicious of everything. The cost of that suspicion is partially yours to absorb. Clear, consistent, paranoia-friendly communication patterns are not just good UX anymore; they are a form of consumer protection.
The most honest thing to be said about AI-assisted scams in 2025 is that we built a very impressive set of tools for generating convincing text, distributed them globally at effectively zero cost, and are now genuinely surprised that the convincing-text market includes people who want to convince you to hand over your life savings. The technology is neutral, sure. But "neutral" is doing a lot of heavy lifting in that sentence, carrying the weight of every grandparent who picked up the phone at the wrong moment.
Stay skeptical. Verify everything. And the next time someone tells you AI is going to make the world dramatically more productive, ask them to be more specific about which world.