Have you ever heard of Deepfake? This term refers to realistic looking face videos. They are videos where someone’s face is made to look like someone else’s. It gained popularity when it was used to add famous actresses into pornographic scenes around 2017 when a forum in reddit named deepfakes was created and it featured face-swapped actors.
Here is an example of a deepfake video of Tom Cruise that went viral on TikTok accumulating around 100 million views:
Advances in machine learning and AI have now made it possible to make it look like a person did or said something. This technology is getting better every day, capable of creating more sophisticated videos every time.
How does it work?
To sum up the process in a simple manner, it’s generated by AI neural networks; deep fake creators use artificial intelligence and a machine learning algorithm to imitate characteristic of real humans. First the computer is fed with real data (images), then the algorithm uses spatial and temporal information to learn the mapping from one person to another.
How to spot a deep fake video?
The difference with traditional fake media is that it’s extremely hard to identify. Deep fake videos or audio have the potential to generate dagame, from political tension to criminal activity. Law makers and tech companies are working on deepfake-fighting technologies but they still have a long way to go.
Here are some things to look for:
- Lower quality section in the same video
- Cropped effects around the eyes, mouth and neck
- The person doesn’t blink or blinks to often in an unnatural way
- Inconsistent skin tone
- Unnatural movements
- Changes in the background
- Unnatural lighting or changes in light during the video
With cyber-security threats on the rise, our experts work to ensure the safety of your information. Get in touch today!