Don’t Get Duped by Deepfakes!

Scammers are often among the first to adopt new technologies so that they can stay ahead of the curve. One such emerging technology is Artificial Intelligence (AI) which is taking over the world by storm.

One of the applications of AI is to create ‘deepfakes’ and as all kinds of media from voices to videos can be ‘deepfaked’, this technology has the potential to cause massive damage and losses when misused. Read on to know what deepfake technology is and how to avoid deepfake scams.

What are Deepfakes?

Image manipulation using tools such as Photoshop has long been an important knife in scammers’ drawers. Image editing is commonly used by scammers to create fake identities, fake documents, false celebrity endorsements and more. Deepfake technology enables scammers to take fakery to the next level by creating videos of people saying and doing things they never said or did. Deepfakes can be created using images, videos and audio too. 

Deepfakes are called so as they were made possible by a branch of ‘deep learning’ called Generative Adversarial Networks (GANs). GANs are based on ‘neural network’ technology but while neural networks can only be used to perceive vast amounts of data, GANs gave neural networks the ability to create fresh content.

Deepfakes have propagated thanks to the machine learning community’s open-source spirit. Advances in generative modelling research are generally made available for free for anyone in the world to download and make use of. As a result, the number of deepfakes have exploded over the years and governments are scrambling to formulate regulations, fearing that the extremely believable fakes can negatively impact society. 

Deepfake Scams

Any technology that has the potential to change the world can be used for good or for evil. Deepfakes have a myriad of positive applications such as bringing the deceased “back to life”, better face-swapping in the film industry, interactive education, immersive marketing and more. Sadly, there’s also a dark side which includes a number of different deepfake scams:

• Impersonation: Deepfakes make impersonating other people easier than ever. In 2019, a voice deepfake was used to scam a CEO out of $243,000. The CEO of a UK-based energy firm believed he was on the phone with the CEO of the firm’s German parent company when he followed the orders to immediately transfer €220,000 to the bank account of a Hungarian supplier. The voice actually belonged to a fraudster using AI voice technology to spoof the German chief executive.
  
  In another case, an unknown user created a fake TikTok account for actor Robert Pattinson who does not have any social media accounts. The account gained more than 660,000 followers and millions of likes before mysteriously disappearing.
  
  
• Catfishing: Catfishing scams involve scammers adopting fake identities using images of random people or celebrities. One of the easier ways to bust a catfish was to ask them to appear on video, as this was not something that could be faked. This has changed with the existence of deepfakes, as catfish scammers can now create convincing videos of the person they are impersonating containing personalised messages.
  
  
• Fake Endorsements: As we revealed in our Global State of Scams 2022 Report, investment scams are now the most rampant scam on the internet. Cybercriminals have begun using deepfakes to promote fraudulent schemes, such as the ”BitVex” platform which was “endorsed” by a deepfake of Elon Musk. They advertised the scam using legitimate interviews modified with deepfake technology to use Musk’s voice in a script provided by the scammers.
  
  
• Deepfake Porn and Blackmail: Porn is still a major taboo in most societies and can lead to anything from embarrassment to social exclusion for people featured in pornographic videos. There have been reports of rejected men creating deepfakes of women to blackmail them. Deepfake porn is also a major concern for celebrities whose faces have been used without consent to create videos of them engaging in sex acts. 

How to Spot Deepfakes

In the early days of deepfakes, spotting a deepfake was as easy as observing if the eyes blink. Unblinking eyes signalled a deepfake as the technology was not advanced enough to create blinking eyes. As the technology improves, it is becoming more difficult to identify a deepfake. However, there are still some signs that you can watch out for by paying close attention:

• Bad lip syncing
• Unnatural blinking
• Stuttering movement
• Missing shadows and unnatural lighting
• Unmoving hair or unnatural hair movement

Generally speaking, deepfakes tend to have the ‘uncanny valley’ effect where the brain senses that something is unnatural. To fight fire with fire, there are tools being developed that use artificial intelligence to identify deepfakes. Deepfake technology is rapidly improving and it is therefore imperative to stay alert and watch out for warning signs of deepfake scams.