Deepfake Technology: Definition and How to Recognize It
Around 300 years ago, a shepherd boy kept crying “wolf” to trick his neighbors, who hurried for rescue from what they thought was a wild wolf attack. One day, a wolf actually appeared, but when the boy cried “wolf,” the neighbors thought this was just another false alarm, and his sheep got eaten.
Today, we are living a similar story. When exposed to news that provokes action or emotion, we aren’t sure whether to believe them or not. However, the difference is that believing them might not be a safe prank. Sometimes, it can cost you a lot more.
So, what should you know about Deepfake technology?
Deepfake Technology Is…
An advanced form of artificial intelligence, named deep learning, allows you to slip the image or voice of any real person in a scenario they were never actually part of. Just like anyone can speak any lie, any person can fabricate any content from scratch and mix any unrelated audio, images or video, to create fake news and information.
How Deepfakes Work
Deepfakes started online in 2017 when someone on Reddit shared fake videos that used the celebrities’ faces on people performing inappropriate behavior. Today, most Deepfake content descends from this same code.
But simply put, Deepfake technology extracts the features of a real person from their photos in different angles and lighting conditions. It then combines them with computer-graphic techniques to generate an image that looks convincing and real, except that it is not. By doing this for every frame, a Deepfake video is created. Because celebrities and public figures have a large count of their photos online, we find them popular victims of Deepfakes.
Deepfakes & Human Rights
The Gulf Centre for Human Rights (GCHR) voiced its concerns about the rising threats of Deepfakes on human rights defenders and their online activism, especially when they support important causes across the Middle East.
Such threats branch into different aspects of human rights as well. For example, Deepfakes can violate your:
- Right to health, when fraud doctors give you ill advice or sell you harmful products
- Right to be free from prejudice, when rumors make you mistakenly judge people
- Right to education, when fake teachers tell you to do things the wrong way
- Right to liberty and happiness, when deceiving news control you in the opposite direction
Deepfakes & Cyber Security
Another major concern about Deepfakes is cybersecurity. Consider the numerous methods currently used for cyber-attacks, and imagine them level up with Deepfakes.
An article in Forbes talked about how in 2019, the CEO of a UK-based agency was scammed out of $243,000 by a fraudster who called pretending to be his boss. The CEO later said that he recognized the accent and the melody of his boss’s voice, although it wasn’t actually his boss over the phone but an AI-altered voice.
How Deepfakes Might Harm You
With Deepfakes, faces can be swapped or created from scratch. Body movement can be replaced. Voices can be faked. Simply, anyone can say anything in any way.
On one side, this means you are more likely to interact with fake accounts or experience deception attempts from others.
On another, others can pretend to be you, for instance, to commit illegal behavior, obtain confidential information on your behalf or blackmail you with reputation-damaging Deepfakes.
How to Identify Deepfakes
Cybersecurity companies are developing algorithms and anti-fake technology to identify any possible inconsistencies that can happen during the faking process. However, you can use your eyes to spot common distortions among Deepfakes:
- Jerky movement
- Flickering on the edges of a face
- Strange blinking or no blinking
- Lips not synced to audio
- Changes in skin tone
- Inconsistent lighting between frames
- Poor rendering of hair, teeth, jewelry, and other fine details
Protect Yourself & Others
Because Deepfake technology keeps expanding, fakes can get more convincing and harder to spot with the naked eye. It’s recommended that you:
- Learn and educate others about Deepfakes
- Keep a skeptical attitude towards what you see and hear
- Make sure you use trusted sources to read the news
- Use online tools, such as Deepware, to double-check the content you suspect
Most importantly, think twice before sharing any content, especially if it provokes negatively intense emotions or immediate action. After all, you don’t want to be a passive partner in crime!