TECHNOLOGY

What is a deepfake and how is it created?

Last updated:

A deepfake is a fake video or audio recording created using artificial intelligence to make someone appear to say or do something they never actually did. It is created by training AI algorithms on lots of real images or videos of a person, then using that data to generate new fake content that looks realistic.

Continue in Reels Listen and swipe through more answers in Technology
Technology UsedDeep learning neural networks, specifically a technique called generative adversarial networks (GANs)
Main Input RequiredLarge collection of real photos or video footage of the target person
Most Common TypeFacial swap videos where one person's face is digitally placed onto another person's body
Processing TimeCan range from hours to weeks depending on quality desired and computing power available
Detection DifficultyBecomes increasingly harder to spot as technology improves, though experts can often identify telltale signs

What is a Deepfake

A deepfake is synthetic media created using artificial intelligence where a person in an existing video or image is replaced with someone else's likeness. The term comes from combining 'deep learning' and 'fake.' Deepfakes can show people doing or saying things they never actually did, making them look and sound completely realistic. They can be created of anyone with enough available photos or video footage.

How Deepfakes Are Created

Creating a deepfake starts by gathering hundreds or thousands of images and video clips of the target person. This training data is fed into an AI system that learns to recognize and understand the unique features of that person's face, voice, and movements. The AI system uses deep learning algorithms to create a detailed digital model of the person. Once trained, the AI can then generate new video or audio that realistically portrays the target person doing or saying new things.

Common Tools and Software

Several software programs and online platforms can be used to create deepfakes. Some are free and available to the public, while others require technical knowledge to operate. Popular tools include FaceSwap, DeepFaceLab, and various mobile apps. The accessibility of these tools has made deepfake creation easier for more people, though professional-quality deepfakes still require significant technical skill and computing power.

How to Spot a Deepfake

Even though deepfakes can look very realistic, there are often small clues that reveal they are fake. Common signs include unnatural blinking patterns, odd lighting that doesn't match the background, unusual skin texture, audio that doesn't sync perfectly with mouth movements, and strange hair or clothing artifacts. As technology improves, these signs become subtler, but experts continue to develop new detection methods.

Risks and Concerns

Deepfakes can be used to spread false information, damage someone's reputation, create non-consensual intimate videos, commit fraud, or interfere with elections. When people see deepfake videos, they may believe false information is true, even if experts later prove it is fake. This technology poses serious risks to society, which is why many countries and platforms are developing rules and detection tools to combat harmful deepfakes.

Sources

  1. wikipedia.org (wikipedia.org)
  2. mit.edu (mit.edu)
  3. nist.gov (nist.gov)
  4. bbc.com (bbc.com)
  5. britannica.com (britannica.com)