The filters are breaking. At least the ones I trust are. For the last couple of years I have watched a steady drip of fake emails slip past Gmail’s defenses. They look nice. They smell like praise. Most of the time they are from nonexistent book clubs telling me my writing is genius, if I only pay their steep fee.
Sometimes the club exists. The emailer just does not. Every author I know is getting them. It is a plague of flattery wrapped in a bill.
I cannot prove AI wrote them. McAfee experts say the odds are high, though. They use the tech for speed. Scale. Personalization.
“AI-powered tactics are making scams more convincing,” said Abhishek Karnik of McAfee. He knows deepfakes, smishing, the fear porn that exploits trust. It makes fraud harder to see.
The danger is real. Click a bad link or open a poisoned attachment. Your devices break. Identity vanishes. Money disappears. You are left with the bill and the stress.
The Mechanics of Deceit
Generators write fast. Machine learning and natural language processing churn out text in seconds. The software scrapes your words. It uses your own blurb to pretend the sender is a fan. It feels real because it is your words just shuffled around.
The scam is slow. They do not ask for money upfront. They wait. You respond. You engage. You feel invested. Then the link appears. Then the attachment drops. The personalization bypasses the filters. The delay bypasses suspicion.
Taylor Peltzman sees it differently. AI did not change the goal. Just the efficiency.
“Organizations are more likely to trust liens from widely used platforms,” Peltzman noted. AI can rewrite messages constantly. It builds multi-step attacks using Google Drive or Microsoft 369. Legitimacy by association is a dirty trick.
The Human Factor
Scams rely on psychology. Fear. Urgency. Trust. AI just turns the crank faster. The old rules apply but they feel heavier now.
Imposters are everywhere. They pretend to be distressed relatives. Banks. Law enforcement. Niche experts in your field. The new terror is voice cloning. Deepfakes can mimic a face. A voice. A writing style. Peltzman warned that even imperfect clones sound convincing if the context fits.
Do not click. Verify the person first.
Shopping scams pop up too. Fake sites. Fake reviews. Unbelievable discounts. AI builds these websites overnight. The products do not exist. The urgency is manufactured.
Pig butchering is worse. They build a small win. You invest in crypto or similar schemes. The AI predicts your risk tolerance. It calculates how much you can lose before you wake up. Then the scammer vanishes with the funds.
Romance scammers have time. Weeks of chats. Months of trust. AI crafts personalized messages for every target. They optimize the love letter. When you fall, they ask for money. For a flight home. For an emergency. They vanish.
Payment apps send fake invoices. Delivery services send alerts for packages you do not remember ordering. Lottery emails claim you won big but need a processing fee. Charities ask for donations using styles scraped from public posts. All of it is easier now. All of it looks legitimate.
The volume is high. The quality is better. That is the problem.
We are all targets now. Not because we are weak but because the noise is loud. Pause. Check. Breathe. The next email might be real. Or it might be a script designed to steal your wallet.
It is hard to tell the difference anymore. Maybe we should just stop replying. But then how do we hear the good news?
