How to detect AI deepfake scams

It looks like the real person, it sounds like a real person, but is real? 

We’re all familiar with deepfakes and we might even have the skills to spot a few. But with technology advancing by the minute, AI deepfakes are fooling even the most educated and confident people as they grow in sophistication. 

How do they work? 

Deepfake technology can mimic not only a person’s appearance but also their voice. This technology has become even more accessible, cheap to use (sometimes even free) and require little to no technical skills. 

With just 30 seconds of audio and video, scammers can make convincing deepfakes that look exactly like the real person they want to impersonate. 

This has made scams using AI deepfakes even more prevalent and harder to spot, where scammers impersonate someone trustworthy, like a boss, family member, or celebrity.  

Is it a real person or deepfake? 

While nine out of 10 Australians say they can confidently spot an AI-generated scam, research from CommBank’s Behavioural Science team has found that Australians can only correctly identify AI scams 42% of the time – which is a lower percentage than a random guess. 

And for those that might think it’s Australians ages 65+ who are being fooled, they’re only 6% less accurate than those younger than them. In this case, age isn’t significant a factor and it shows that deepfakes can fool anyone, no matter the age.  

This overconfidence factor is certainly dangerous in the world of advanced AI deepfake scams, where Australians are unlikely to stop, check, and reject when it comes to making payments and transfers under pressure. 

Scammers know that humans will trust familiar faces and voices, and so, deepfakes are the perfect tool for manipulation. You might be reading this and think, that wouldn’t be me. However, these deepfakes are so convincing, the difference can be (literally) by a hair!  

Protecting yourself from deepfakes  

Nevertheless, knowing that AI deepfakes are advancing by the minute, we shouldn’t be scared or overcome by them. Instead, we ought to be vigilant when we’re asked to make any transaction under pressure, even if it’s by someone you trust. 

Here are some of the ways to practically protect yourself from AI deepfake scams: 

Scammers use deepfakes of well-known experts and celebrities to promote fake investment opportunities. It’s important to speak with someone you trust like your independent financial advisor before transferring money and check ASIC’s Moneysmart Investor Alert List. 

Scammers can use voice cloning technology and deepfakes of loved ones to gain trust and manipulate people. Set up a safe word that only your family knows to protect each other. 

Scammers can create real-time face swaps that appear authentic to pretend to be a potential love interest in photos and video calls. 

Never send money to a romantic interest if you’ve never met them in person. Remember, even video calls can be faked. It’s important to talk to a friend – secrecy is a major risk factor – and arrange to meet them safely in person. 

Lastly, approach any transfer or payment under pressure with caution – the same principles of “Stop. Check. Reject” should still be applied. 

Slow down, check the details, speak with someone you trust, and if it’s a scam, stop contact with the scammer and block all contact with them. 

For more information and videos on AI scams, search “CommBank Safe”.