Are you being fooled? The truth about AI scams you walk into every day

Updated March 10, 2026, 8:24 a.m. | By Carol Ofori

Scammers are now using the same AI tools you use at work. This is how to tell the difference between a real website, call or chatbot and a fraudulent one designed to steal from you.

A humanoid robot has a thinking pose/iStock/iLexx

The same technology used to build legitimate websites, write reports and assist with everyday tasks is now being used by fraudsters to construct elaborate, believable deceptions.

According to My Broadband, deepfake videos of public figures, including news presenters and heads of state, are virtually indistinguishable from genuine footage.
Voice cloning has become a particularly dangerous threat. In one bank scam, fraudsters cloned the voices of real bank employees to call customers with urgent "security alerts."

The calls came with spoofed caller IDs and sounded entirely authentic, but the red flag was clear: they pressured customers to share a one-time PIN or approve a transaction. No legitimate bank would ever do that.

ALSO READ: Hearing test on WhatsApp? This SA university is breaking the mould in hearing health education

Why are AI scams so difficult to spot?

AI-powered scams remove all the traditional warning signs. Spelling and grammar are flawless, branding looks convincing and fabricated customer reviews alongside phoney customer service replies make fake e-commerce sites appear entirely credible.

Back-to-school scams have illustrated just how sophisticated these schemes have become, with fraudsters building fake school registration sites, counterfeit stationery and uniform shops and AI-generated bursary and scholarship posts that looked completely legitimate.

What is the difference between a legitimate bot and a scam bot?

Not all bots are malicious. In the right hands, AI-powered bots are useful customer service tools that provide round-the-clock assistance.

However, scammers are now deploying their own bots to conduct convincing conversations designed to defraud, typically through romance scams, fake investment schemes or phoney customer support.

ALSO READ: A private school for R780 a month? Here's all we know

How do you identify an AI bot scam?

There are specific patterns that reveal a scam bot. Responses arrive instantly at any hour. The language is unusually formal and conversations feel generic.

Bots give vague answers to direct or personal questions, frequently repeat phrases and push users urgently towards links, payments or specific actions.

If a public figure appears to be announcing something shocking that demands an immediate response, treat it as fake until you have independently verified it is real.