What can we believe on the Internet? The dangers of Deepfakes
Would you fact-check the content that you see on the internet? Most people do not, However, there are a huge amount of misinformation and even disinformation blooming on the internet…so what can we believe?
What is “deepfake”? The combination of “deep learning” and “fake”
The term “deepfake” comes from “deep learning” and “fake”. The technology of artificial intelligence (AI) deep learning can train themselves based on given data (images or videos) which are used to change faces in video or any digital content, and finally, synthetic media (often are video, image and audio) will be generated.
The US media Buzzfeed and filmmaker Jordan Peele co-produced a deepfake video in 2018, to raise public awareness about deepfakes.
In short, deepfake tech uses AI to portray someone saying and doing things that never happened, meaning that you might watch a person in a video, but that is not that person, it is an AI creation in the person’s likeness. Moreover, a deepfake video often includes someone’s image or voice but without their agreement.
In addition, deepfake needs adequate images or videos for AI to train themselves to look authentic. Thus, celebrities and politicians whose photos are exposed on the Internet, countlessly, will be the target of deepfakes.
Another tech that needs to be mentioned is Generative Adversarial Network( GAN) which would improve the quality of the creation of deepfake, making fake content more real.
You could see some creations made by GAN on this website, This Person Does Not Exist.
Social media make deepfakes more rampant
Social media, for example, Instagram, Facebook, Twitter and so forth provide open platforms for users to share things. These platforms are helpful for the freedom of speech, however, also provoke a concern that people could share content that is made by deepfake tech, easily.
Social media platforms where rumours, disinformation and misinformation spread easily, people might be manipulated by the phoney content if people do not have awareness about deepfake.
Deepfake in Taiwan: YouTuber’s scandal
Taiwanese YouTuber Hsiao Yu who had about 1.3 million followers was arrested for using deepfake tech to put celebrities’ faces on actor’s in pornographic videos.
According to Radio Taiwan International (RTI), the deepfake videos he made has profited him around NT$10 million, totally. The victims were up to 80 people, including celebrities and politicians
As reported by CNN business, 96% of the deepfakes are made for pornographic content, other than that all of the content featured women. It is evident that deepfakes tech has been used for filthy things, and the trauma that victims had are hardly be recovered.
Making the dead alive? Deepfake involved in the movie industry
On the other hand, deepfakes can be used, positively. An American actor, Paul Walker who had died during the production of Furious 7, a famous American action film. However, deepfake tech brought him back to the screen.
Paul’s brothers, Caleb and Cody, who have a high likeness to Paul are involved in the production of the film, helping this movie to be completed. As a result, deepfake has given the movie industry more possible than we might think.
Social media have the responsibility to regulate the misuse of deepfakes
In fact, social media platforms have moderated the abuse of deepfakes. For instance, Facebook has banned the creation of deepfakes in 2020; Twitter, on the other hand, enforce every deepfakes video need to be labelled as synthetic but taken down if the content is likely to harm people.
However, the regulations against deepfakes are not perfect, all in all. Especially, on social media platforms that accelerate the spread of fake content, moreover, the platforms cross, multinationally, but the law is managed in a certain nation.
As reported by Coda Story, social media companies know that they will have huge losses if they execute content moderation, because of user traffic, it is obvious that they need to be given external incentives. As a result, the governments need to intervene, and international cooperation might be essential.
Outside of regulation, people need to raise awareness about the double-sided sword of emerging tech. Governments should not only address the misuse of deepfakes by cooperating globally but also publicize the hidden dangers of deepfakes by joining NGOs, tech corporations, fact-check organizations, and so forth.
Feel free to reply to this article if you have different perspectives or anything you would like to share with me.
References:
- PanSci: Deepfake 不一定是問題，不知道才是大問題！關於 Deepfake，你需要知道的是⋯⋯？
- TIM Review: The Emergence of Deepfake Technology: A Review
- Insider: What is a deepfake? Everything you need to know about the AI-powered fake media
- CNN business: The number of deepfake videos online is spiking. Most are porn
- Weta Digital: Furious 7 VFX | Breakdown — Brian O’Conner | Weta Digital
- Coda Story: It’s not too late to regulate deepfakes
- Newsweek: 深度偽造雙面刃 新科技背後的危機
- BuzzFeedVideo: You Won’t Believe What Obama Says In This Video!
