{"id":96039,"date":"2024-07-25T07:45:42","date_gmt":"2024-07-25T15:45:42","guid":{"rendered":"https:\/\/www.trendmicro.com\/internet-safety?p=96039"},"modified":"2024-08-05T03:06:52","modified_gmt":"2024-08-05T11:06:52","slug":"deepfakes-101","status":"publish","type":"post","link":"https:\/\/www.trendmicro.com\/internet-safety\/blog\/deepfakes-101\/","title":{"rendered":"Deepfakes 101: What are Deepfakes and How to Spot Them\u00a0"},"content":{"rendered":"\n

The term, “deepfake<\/strong>“, is a blend of “deep learning” and “fake”, and was coined in 2017 by a Reddit user. A deepfake is a fake image, video, or voice recording where one person\u2019s likeness is replaced with another’s. These are made to deceive or entertain people. With advanced machine learning, deepfakes can look very realistic. Previously, deepfakes were primarily known for their humorous use on social media. However, their potential for abuse quickly became apparent and deepfakes are now a significant concern for privacy, security, and information. In a recent Trend Micro study conducted in June 2024, 71% of people reported they feel negatively about deepfakes<\/strong>, with a belief that the top reason anyone creates deepfakes is for fraud.1<\/strong><\/sup> <\/p>\n\n\n\n

How Are Deepfakes Made? <\/h2>\n\n\n\n

Creating deepfakes involves using advanced computer programs called Generative Adversarial Networks (GANs) to make fake images that look real. There are four main steps: <\/p>\n\n\n\n

    \n
  1. Data Collection<\/strong>: Firstly, media content (images, videos, or audios) of the target person is gathered. This collection is used to train the computer program. 
    <\/li>\n\n\n\n
  2. Training the Model<\/strong>: The GAN is trained with the collected data. One part of the GAN creates fake images, while the other part checks if they look real. 
    <\/li>\n\n\n\n
  3. Refinement<\/strong>: Techniques like facial landmarks and motion capture are used to make the deepfake look natural, with realistic expressions and movements. 
    <\/li>\n\n\n\n
  4. Final Production<\/strong>: The finished deepfake is then combined with the original media, creating a seamless and convincing piece of fake content. <\/li>\n<\/ol>\n\n\n\n

    While the above may sound somewhat complex, the fact is anyone can make a deepfake, given the huge number of software applications accessible to the public, from DeepFaceLab to DALL-E<\/strong> and Midjourney (though the latter has safeguards in place). <\/p>\n\n\n\n

    \"Illustration<\/figure>\n\n\n\n

    <\/p>\n\n\n\n

    Why Care About Deepfakes?<\/h2>\n\n\n\n

    It is easy to think of cybersecurity as an abstract concept, removed from everyday life \u2014 but the implications of malicious deepfake use affect the individual and society at large: <\/p>\n\n\n\n