What Is a Deepfake? Is It Legal? Everything You Need To Know
Computers are getting smarter by the day. From movie graphics to AI-powered tools, almost everything can be simulated. Deepfake technology has now been making the headlines. So, what is Deepfake and is it legal? Read on to learn more.
What Is Deepfake?
A deepfake is created when AI replaces one person with another in a video. The term comes from deep learning. It is similar to Deep learning algorithms that teach the computer how to solve problems with huge data sets. This algorithm is used to swap faces within a video or digital content and make it realistic.
How Are Deepfakes Created?
There are many methods to create deepfakes. However, the most common is with the help of deep neural networks that use auto-encoders to swap faces. You will first need a target video and a collection of various clips of the person that you want to insert in the video. Both videos don’t need to be related. Auto-encoder is an AI program that studies the video clips and understands how the person looks for different angles and conditions. It then maps that person into the video by finding various common features.
GANs, or Generative Adversarial Networks, are also added to this. It is a type of machine learning that improves any defects in the deepfake video and makes it difficult for anyone to decode it.
How To Create Deepfakes?
There are various apps available to create deepfakes. Various deepfake software can also be found in the open source community. While some are for entertainment, others are more malicious. Experts believe that the deepfake technology will become even more dangerous and may pose threats to the public.
Are Deepfakes Only Limited To Videos?
Deepfakes are not just videos, there are various other uses of it as well. For example, realistic audio deepfakes can also be made. This is similar to video cloning, where deep learning algorithms study the tone and depth of a voice and clones it into an audio recording.
How To Stay Safe From Deepfakes?
Deepfakes are getting more and more realistic. However, there are a few defensive techniques. Various companies are developing methods to spot a deepfake. While this is being developed, you should be aware of how to spot a deepfake without the use of technology. Check if the person in the video is not blinking or blinking too much.
The current deepfake technology has not perfected animating faces and this results in the person not blinking or blinking too much. You should also look for defects with skin or hair or the face seeming blurrier than the background.
Also, check the lighting. Does it look unnatural? Is the background brighter than the person? More importantly, if you detect something against a person you know, make them aware of it. Deepfakes have the potential to cause a lot of damage if unchecked.
Advertisement
What is it with these click-bait headlines?
When I see titles with phrases like “Everything You Need To Know…” or “Everything We’ve Heard About …” or “The 5 Things……” , etc., almost invariably the actual articles are shallow, simplistic, consist of only 5 or 10 short paragraphs and not worth reading.
There is seldom any intellectual or educational nutrition to these kind of stories. Seemingly, they are written for the lowest common denominator – or only to create traffic statistic the web page can present to get advertisers to pay them for placing ads?
Oh, ghacks. Finding good content here is getting harder every day…. ?
I know two types of deepfakes
1) Face deepfake. Used in porn. Not a fan of deepfake porn.
2) Voice deepfake. Used in extortions and other money scams. Extremely effective.
Those are also used for organized harassment of individuals by state fascists. There is really lots of damage that can be done when such a powerful organization can substitute itself invisibly to everyone you’re contacting, especially if done subtly and over several years, and in combination with other tools.
Time for some new laws to be invented.
We’ve been in the “Kleenex” (change, don’t keep) era for some time now. Here we come up to the “Fake” epoch.
Deep, deeply fake, fake, fake. What is the point? Spending time to find out if we’re facing authenticity or fake after smart guys had fun mixing up reality?
AI, Deepfake and Company : count me out. AI for science, technology : great. Not otherwise, not delivered to the masses, not used by corporations to further enslave the masses.
Basically ‘fake’ is nothing but a lie. “You shall not lie” are we told. Told but no one listens to morality any more. OK, call it ethics, but a lie remains a lie and ethics won’t deny that as it won’t deny lying is not only ‘bad” but a fundamental mistake, a human error. We all happen to lie but when lying is elevated to the rank of normality then our world is facing big problems.
May I be crud? Deepfake is bullshit (just to prove I’m not anchored to a morality fundamentalism).
+1
“Is It Legal?” Did I miss the answer to this Headline question?