Deep Fakes in the Entertainment Industry

Artificial intelligence can create deepfakes, which are synthetic media that replace a person's likeness in an image or video. Deepfake technology can deceive viewers by showing someone doing or saying things they didn't do. It is based on deep learning and can be used for various purposes, including entertainment. An example is a deepfake advertisement by UltraV company to raise awareness of toxic beauty advice. Deepfake technology has educational value in this ad as it makes teenagers question the authenticity of such advice. However, there are ethical concerns with deepfake technology, including the creation of non-consensual pornography and widespread mistrust in society. Hello everyone, welcome to the Artificial Intelligence podcast. I'm Li Junyao, and today we will explore how artificial intelligence can be utilized to create deepfakes in the entertainment industry, and the possible ethical issues that might be involved. Before we start, I'm delighted to welcome a guest, Mo, and she will also participate in this podcast today to provide her insights into the particular use of deepfake technology. Well first, let's talk about what deepfake technology is. Deepfake is a type of synthetic media in which a person in an existing image or video is replaced with someone else's likeness. Such technology also can simulate a person's voice. So deepfake can deceive the viewer by showing someone's likeness in the video, but doing things or saying things that they didn't do or say. The impacts of deepfake are based on deep learning, and involve training artificial neural network architectures. The more relevant training data used in deep learning, the more synthetic video and audio will be realistic. That's why well-known politicians and celebrities have often been subject to deepfake technology. Because deepfake technology has become more convincing and accessible to the public, it has been used for various purposes, but mainly for the entertainment industry. Here is an example of deepfakes being used by a major brand and for a marketing activity. With a positive sign, UltraV company created an advertisement entitled Toxin Influence. The aim is to raise awareness of the negative impact of the dangerous beauty's advice on social media apps. Most teenagers are invited to watch videos produced by the company. In the video, this company used deepfake to make teenagers' mothers say some ridiculous advice about dangerous beauty. Like Botox is amazing, you are never too young to start. After watching this video, these teenagers consider that, although this content about toxicity can be seen on social media every day, they were still horrified by the distorted advice given by their mother. In order to explore the impact of using deepfake technology in this advertisement, I invited my friend Mo to watch this campaign video. Mo, what do you think about the deepfake video in this advertisement? It's pretty intriguing. It looks so real. The facial expressions of the moms in the video are lively and vivid. So it's hard to tell if AI, especially when you didn't tell me this technology is used before watching the video. Which makes the whole thing quite scary though, because I would imagine this happened to me in real life. And I might be as shocked as the girls. Okay. And what do you think about the value of using deepfake in this activity? I think using deepfake has some educational meaning in this ad. These girls would see toxic content online on a daily basis, but they never realized that they were subconsciously brainwashed and that they are so directly influenced. And when these toxic comments are coming out from their closest ones who love them the most, these girls would doubt the authenticity of the comments instead of themselves in the first place. So I would say that using deepfake technology in this ad definitely works to raise viewers' awareness of toxic beauty advice online. Yes, I agree with your opinion. The video with deepfake technology is instructive. It could help parents navigate tough conversations and also allow parents to have a channel to guide their children, allowing teenagers to reflect and have a more positive experience on social media. Nowadays, deepfake is not limited to commercials and films in the entertainment industry. Such technology is beginning to be accepted by the public. For example, face blocking pictures is the new trend on social media from matching a specific user's facial expression in real-time to generate face-blocking videos. In fact, deepfake technology traditionally needs to be trained with lots of video footage of this person, including a wide range of facial expressions and under all kinds of different lighting. However, as deepfake technology develops more sophisticated, creating a deepfake video is much faster than a couple of years ago. Even some apps can swap your face with another person in a photo by using only a handful of images, which is one. In order to better understand how people can easily utilize the technology of deepfake in our daily life, I invited my friend Mo to use a single-photo face-blocking app called Reface. I uploaded our recent photos on Reface to change Mo's face onto mine. Hey Mo, how do you feel about this synthetic picture? It's pretty cool. Although there are some small differences, it looks real in general. It's astonishing how far this AI technology has gone. You can simulate a person based off a photo now, but does it mean that we can swap faces with anyone on this? I think the terms of service and privacy policies on the app don't have any explicit restrictions. In this app, you can even put your own face on any politicians and celebrities. Wow, then this could be risky. It means that anyone can use my photo without my permission. Aren't there any ethical issues involved in apps using deepfake technology? I mean, I'd assume a lot of users would concern about this. Yes, there are some ethical issues involved in using deepfake technology. In this case, anyone can do facial swapping by using free apps, eliminating the need for technical abilities. The simplicity of deepfake magnifies its violations of the right to portrait and privacy. To be specific, one ethical concern surrounding deepfake technology is its potential to be used to create no-consensual pornography. Deepfake has already been used to create fake pornographic videos using the face of real people without their consent. This is a serious violation of privacy and can cause significant harm to the individuals involved. According to the research of Sensitive AI Company, between 90% and 95% of all online deepfake videos are no-consensual porn, and around 90% of those features women. Besides, DeepNudes can be an example. This is a nude picture generation app that was launched in 2019. This application creates nude images of women by only uploading one picture. These shared photos will be shiftily spread on the internet without the victims realizing their privacy has been violated. As a result, the victims' reputation and mental health will be suffered considerably. Moreover, deepfake pornography can be used for blackmail or extortion. Victims will be threatened to release compromising videos or images unless they pay a ransom. If we can't tell whether it's true or fake, how are we gonna identify truth from all the fuss we see on the internet? And I think it would be even trickier for younger people because it obviously requires more critical thinking. Yes, it is another ethical issue I want to mention. Deepfake technology would cause widespread mistrust in society. This has already happened. For example, a fake video of Ukraine's president's security. In that video, the president orders the surrenders of the national army. The video went viral online, but was debunked as a deepfake. However, when the public realizes that deepfake is being applied on a broad scale and they have no way of discriminating whether it is realistic, they will start distrusting some official communication channels such as news media. In this example, deepfakes are deliberately deployed to incite social conflict on the internet, which would lead to a serious consequence. And some researchers illustrate that the improper use of deepfake blurs the borders between fact and fabrication. Such damages caused by deepfakes are often long-lasting and irreversible. Okay, Mo, thank you so much for joining me to explore the application of deepfake and its potential ethical issues. That is all for this podcast. Thanks for listening.