It's Not Just Deepfakes Anymore
When people picture AI-driven scams, they usually think of a cloned voice or a fake video call. That's real and growing, but it's only one piece of a much bigger toolkit. AI now helps scammers write phishing emails with none of the grammatical tells that used to give them away, generate entire fake e-commerce sites in minutes, produce fake product reviews at scale, and personalise mass scam messages using details scraped from social media and old data breaches - all without needing any real technical skill. The barrier to running a convincing scam has dropped dramatically, which is a large part of why fraud volumes have climbed so quickly.
The Scams Worth Actually Knowing About
Romance scams increasingly use AI-generated photos of people who don't exist, paired with AI chat assistance that keeps a convincing, always-available conversation going for weeks before the money request arrives. Investment scams use deepfake video of celebrities or financial personalities "endorsing" a platform, often on social media, where reach is highest and scepticism is lowest. Job scams use AI-written listings and AI-run "interviews" to collect personal information or upfront "training fees" from people who are genuinely job-hunting. Tech support scams use AI voice systems that sound exactly like a real call centre, right down to hold music and a queue number.
Why These Work Better Than Older Scams
Older scams often had a tell - broken English, an obviously fake logo, a website that loaded slowly and looked amateurish. AI has quietly removed most of those tells. Grammar is now flawless. Logos and layouts can be generated to match a real brand convincingly. A chatbot can hold a natural, responsive conversation instead of sending an obviously scripted message. The psychological playbook hasn't changed - urgency, authority, emotional pressure, a request that something be handled quickly and quietly - but the surface-level cues people were trained to spot have largely disappeared.
What Still Works as a Defence
The good news is that the underlying manipulation tactics haven't actually changed, so the underlying defences still work. Any message that combines urgency with a request for money, gift cards, cryptocurrency, or remote access to your device is a red flag regardless of how professional it looks. Independently verify - go to the company's real website by typing the address yourself rather than clicking a link, or call a number you already have rather than one provided in the message. Be sceptical of relationships or investment opportunities that developed unusually fast, especially ones you can't verify through any channel outside the platform where you met. And treat any request to move a conversation off a mainstream platform to a private one - often a scam tactic to avoid detection - as a warning sign in itself.
If You Think You've Been Targeted
Stop all contact and payment immediately, contact your bank if any payment was made, and report the incident to your country's cybercrime or consumer protection authority - reporting helps other people even if recovering the money is unlikely. Scammers rely heavily on victims feeling too embarrassed to report what happened; that embarrassment is exactly what allows the same operations to keep running.
