A highly sophisticated scam – such as creating a fake video – was once so expensive it was only worth the scammers time to target high profile, rich victims. Deepfake videos – where AI creates a film of an individual, or cloned voices – where using AI, someone can use another’s voice to say anything, are so easy and cheap to do, that anyone can be a cost-effective target for a scammer.

The software required to do this has many legitimate purposes and it is easily accessible. While we shouldn’t live in fear, we must recognise that AI makes scams more sophisticated and sadly, more effective. A long, detailed conversation with a loved one is unlikely to be a scam, a short more detached request for help from a loved one’s voice may be. Particularly if that help takes the form of a gift card or cryptocurrency. To avoid being the target of AI scammers, here are a few tips, as well as things to look out for:

1. If it seems too good to be true, it’s likely a scam: as we’re talking about AI scams, it would be remiss not to mention, more traditional financial and investment scams which are leaning into the hype around AI to convince people to part with their money. Fraudsters are leaning into people’s fear of missing out on the AI boom in investments to bombard them with offers of get rich quick schemes. As always, if it seems too good to be true, it is likely a scam, especially if you are asked to pay deposits quickly to secure your ‘opportunity’ or pay in unusual ways such as with cryptocurrencies.

2. Establishing safe words/phrases to combat voice cloning: it is very easy, even with a small sample of someone’s voice, to create a very realistic copy. This lets scammers pretend to be loved ones, business partners and more, and make (usually short) phone calls. In these calls they will usually and quickly ask for help and request you send them money. Scammers prey on vulnerabilities and worry, the call may come in the dead of night with a loved one’s voice explaining they have been in an accident, borrowed a phone and urgently need you to transfer some cash to an account to pay for treatment. Even the most cynical of us may be tricked if the conditions are right. These scams are increasingly common, and people may want to consider establishing safe words or phrases with their family to be used to verify that a caller is who they say they are.

3. Determining if it’s a deepfake video: it is now cheap and easy to create a video where an individual looks as if they are saying or doing something that they did not. You have likely seen videos of celebrities seemingly endorsing various investment scams. These are fake but used by scammers to generate credibility for their scams. Deepfake videos can be used for identity theft and to pass verification checks. They can be used to extort people, with scammers threatening to release deepfake videos of people in compromising situations, or to convince people they are dealing with legitimate businesses and even law enforcement.

While it is worth noting that, as AI improves, so do the deepfakes – and good deepfakes are virtually impossible to spot. However, deepfake videos may contain tells. Always look out for lip-syncing errors, unusual facial expressions or strange backgrounds. The subjects may stay unusually still, appearing rigid and monotone. The subject may blink too often or too little and, if wearing glasses, they may not look right – perhaps too much glare (or too little). AI also often struggles with facial hair, so beards or moustaches may not look quite natural.

4. A new generation of phishing emails: AI makes more traditional scams like phishing far more effective. We are usually used to random emails enticing us to click on a link to claim a prize or similar. However, AI can supercharge these attempts. It is far cheaper to create a convincing looking email or website, virtually indistinguishable from the real thing. The new generation of phishing emails are highly convincing, looking like urgent notifications or genuine requests from your bank or businesses you deal with. They are highly personalised, containing the information needed to make the email look credible. Using information available online, scammers using AI can cheaply create a communication which bypasses many people’s initial scepticism. As convincing as the email might be, something coming unexpectedly and requesting that you give information is cause for concern. If you are not expecting it, check.

5. Spotting the new age signals and taking the time to verify your information: many people are naturally cautious online. They recognise an email claiming to be from their bank, which is littered with spelling mistakes, may be dodgy. People getting a random call from a robotic voice are not likely to pass over their details. However, in this new AI world, all of these signals are a thing of the past. The next phishing email you get may well be very convincing, the robotic voice may now be a loved one calling for help in the middle of the night. Scammers are upping their game with AI. People should be cautious of anything unexpected. Taking your time to verify information prevents scams, scammers will always try to put you off doing that.

6. The red flags to look out for: always be cautious if someone contacts you unexpectedly. If you get a call from your bank asking for your details, hang up and call them back on their official number. This is however far easier to do than if a loved one calls to say they have been in an emergency! Whilst spelling mistakes and poor punctuation in phishing emails may be a thing of the past, also beware of highly formal language which could come from Generative AI. Things may still look a little unnatural. Scammers will try to catch you out by creating a sense of hyper-urgency. They will say that something needs to be done now, there will be no time to check, and checking will make the situation worse. This is designed to force you into making a quick decision without time to check anything out. Something which needs to be done now is often a big red flag for a scam. As are requests for unusual (and unrecoverable) payment forms. Be cautious if someone requests you to pay in things like cryptocurrency or gift cards.

AI is good, but not perfect. If you see slight inconsistencies, glitches or signs something isn’t quite right, trust your instincts. 

With AI, scammers are able to generate an increasing volume of highly sophisticated scams. Always err on the side of caution and remember that scams are often designed to catch you off guard, to not give you time to think, and to pressure you into making a decision now.