Deepfake videos

How deepfakes happen

Deepfakes are one of the most talked-about topics in the world of technology. This technology is a breakthrough solution for many industries, but its negative side has much more impact than the positive. Let’s discuss how deepfakes happen.

There has been a recent rise of concerns about the ethical use of deepfakes and how they can be misused for defamation, scamming, and phishing. So if you’re worried about deepfakes and wondering how to prevent them, you’re at the right place. 

Read on to learn how deepfakes work and how you can use Proofpoint to prevent it. 


What Are Deepfakes?

Deepfaking refers to the application of synthetic media to create speech, images, or videos with someone else’s voice or appearance. This technology has extremely unethical roots, with the first reported use in 2017 by a Reddit user to create face-swapped content. 

More recent uses include the manipulated recorded speeches of public figures such as Mark Zuckerberg and the prime minister of Belgium saying things that could jeopardize their reputation. 

Deepfakes can even be as simple as imitating your texting tone or accessing your online accounts and using your online persona for the hacker’s benefit. However, Deepfakes typically begin with a simple phishing email

deepfakes video
how deepfakes happen

It’s pretty easy for hackers to learn your online habits, so they can send emails that look a lot like your regular, everyday emails. Once you open the email and click on the link, the hacker instantly gains access to your device in multiple different ways. 

They could have access to your email account, social media accounts, or the OS entirely. Then, the hackers meticulously collect the necessary information over a few months, such as your banker, voice recordings, bank information, and more. 

Two of the primary resources they need for their attack are your voice recordings and email account. Phishing allows them to gain access to both with ease. Then, they may email your banker or call them by deepfaking your voice and pretending to be you. 

They repeatedly call the banker to confirm the transaction, making it easier for the banker to believe that it’s you.

how deepfakes happen

How to Prevent Deepfakes

Here are a few measures to take on an individual or organizational basis to prevent deepfakes


Proofpoint is the best way to stop the threat of deepfakes at its root: the email. This industry-leading gateway provides email protection from both known and unknown threats, specializing in phishing and spam. 

Proofpoint’s email protection system is powered by NexusAI, allowing the software to detect and block threats with malicious content instantly. The Advanced BEC defense of this software runs on AI and machine learning, analyzing the header data, IP address, and message body to fish out BEC attacks. 

Whether you’re dealing with phishing, imposter threats, malware, or spam, Proofpoint utilizes multi-layered detection to protect your device against constantly evolving threats. The software essentially acts as a filter to ensure you only receive safe and important emails. 

Other Solutions

Besides using Proofpoint, you must also take manual measures to ensure deepfakes don’t compromise your personal information. Here are some of them:

  • Use semantic passwords for your private conversations to ensure you’re not on call with deepfakes. That can be a code word or a secret question to ask or answer. 
  • Ensure your biometric security feature providers have up-to-date and reliable tools. 
  • Educate your employees about how to detect deepfakes and avoid them in the future. 


Deepfakes can be detrimental to your public image and your financial assets as well. It’s important to know that most deepfaking begins with phishing emails, which is why you need to protect your email account diligently. Reach out to Mac Help Nashville to prevent deepfakes in the future.

Mac Cybersecurity Threats, You May Not Be As Safe As You Think

4 thoughts on “How deepfakes happen”

    1. Nate, first of all this article is probably my shittiest one.
      It is only a rough view of the types of technology that could do this.
      For this example I had to read a few pages of text and you can still hear the robotics in the voice.
      Of course it is much easier to forge a public figures voice because there will be much more samples to ingest for training.
      But as AI (humans teaching and programming) advances and is able to fill in the blanks better and better, vocal simulation will become easier.
      If someone could record the phone calls or gather interview videos or social media videos that soon may be enough.
      Right now we have to choose very carefully our voice generators, is a great very smoothe normal sounding human voice with just typing text.
      The voice models were trained by having the subject read text into a michrophone.
      So as audio engineering and AI begin to cross into each others space, soon, we should be able to fake a voice with just a few sentences.

Leave a Comment

Your email address will not be published. Required fields are marked *