Today, the popular social networking site TikTok hosts a river of fake videos known as “deepfakes”.
A deep fake can digitally alter a person’s face or body to make them look like someone else, usually a famous person.
A prominent TikTok account @deeptomcriuse has already posted dozens of fake videos impersonating Tom Cruise, gaining 3.6 million followers.
Other examples of deepfakes show Mark Zuckerberg – the founder and CEO of social networking site Facebook and its parent company Meta – admitting to conspiring to share users’ data. There are recent examples of actors such as Robert Pattinson and Keanu Reeves.
Although deepfakes are often used to produce creative works or for entertainment, they are increasingly common in disinformation campaigns, identity fraud and in the defamation of public figures and celebrities.
Since the technology to produce these counterfeits is so sophisticated and complex, they are becoming increasingly available, leaving detection programs and laws behind. One thing is certain about it, that deepfakes are here to stay, so what are we doing about it?
According to the report published by The Conversation, the manipulation of texts and images has always been the cornerstone of interaction and attracting attention. Deep counterfeiting is no exception. It reflects people’s deep-rooted desires to participate in culture, storytelling and art.
Deep forgeries can produce original works by deceased actors, as if resurrecting them into new, original works. It can play a role in comforting people who have lost loved ones.
High and tight then… eyebrows ✂️ #footloos
♬ Footloose – Kenny Loggins
Disadvantages of deep fake
Deepfake techniques can cause many problems, including social problems such as:
• Use it as proof of any fake news or misleading information.
• Use it to discredit persons.
• Causes social or political conflict due to its ability to produce media whose authenticity is difficult to trace.
• Its use in the production of fake pornography.
The last point is particularly worrying, as Deeprace’s deepfake detection software showed – in 2019 – that 96% of the total 14,000 cases of deepfake were of a pornographic nature. A now-removed free app called DeepNude 2.0 could reveal women’s clothes to show them naked. This application was either used for the purpose of revenge or to blackmail the victims.
In Australia, some offenders have managed to circumvent “revenge pornography” laws by using deep fakes; This is an issue that is expected to become more serious in the coming days.
In addition, deep fake techniques are used in identity fraud, especially in the form of video messages that reach people pretending to be from a “colleague” or a “family member” asking for money. One study revealed that identity fraud – using digital manipulation – cost US financial institutions $20 billion in 2020.
The makers of deepfake clips claim that it takes a lot of time and effort to make them look realistic. “It’s not just a single click,” Chris Ohm, creator of the fake Tom Cruise account on TikTok, told The Verge.
However, there is new evidence that deepfakes are getting easier. Researchers at the United Nations Global Pulse show that speeches can be faked in just 13 minutes.
As more deep fake apps develop, unskilled people will be able to produce more authentic fakes.
The need for governing legislation
It is worth noting that the Facebook platform was criticized for its failure in 2019 to remove a fake video clip of the US House Speaker Nancy Pelosi, some of the failure of the platform to distinguish it as fake.
In 2020, Twitter banned the sharing of synthetic media that deceives, misleads, or harms people (other than simply adding logos or tags to that media). Tik Tok did the same. YouTube also banned deepfakes related to the 2020 US election.
Even if this policy is good, the response to reports asking for the removal of deepfakes will not be fast enough. And these policies are quickly put into effect when protecting celebrities and politicians, while falling short of protecting ordinary people from deep fakes.
Did you guess any of the moves? #pattinson #neverletthemknowyournextmove #nextmovechallenge
♬ Original Sound – Unreal Robert Pattinson
Deep fake detection
People today need to equip themselves with as many skills as they can detect deep fakes. There are questions whose answers can help you judge it, including:
• Is the face smooth this way? Are there shadows for the cheekbones out of place?
• Do eyelid and mouth movements seem somehow unrelated or abnormal?
• Does the hair – especially facial hair – look fake? It is difficult for current deep fake techniques to maintain the original appearance of hair (especially facial hair).
• Ask yourself about the message the person is sending: for example, is it to avoid vaccinations, or is it a pornographic clip? Anything that appears unusual, normal or contradicts collective knowledge is usually associated with deep fake.
Source : The Conversion + Websites + Social media