If such AI is deployed on a large scale, it will upend the rules of the game for studios, content creators, and producers. This is almost inevitable.
Written by | Slightly fat
In 2020, the Korean film "Parasite" won the Oscars for Best Picture, Best Director, Best Original Screenplay and Best International Film, rewriting the history of cinema.
In his acceptance speech, director Bong Joon-ho said, "Once you overcome the one-inch-high subtitle barrier, you will see more wonderful movies."
"Parasite" won best picture, becoming the first non-English film in Oscar history to win the award.
Long before Parasite, some foreign-language films broke through the subtitle barrier and were successful in the mainstream Hollywood market. But this does not affect the real existence of subtitles as obstacles.
Outside of streaming, international films and TV series still struggle to attract audiences. As of 2020, only 12 movies have been released in more than 1,000 U.S. theaters. Before Parasite, the last one was Jet Li's Fearless in 2006.
In fact, when Parasite first came out, the Academy Award winner was only released in three theaters. As its success demonstrates, a great film should be able to appeal to an audience, regardless of language. However, many award-winning or nominated subtitle films rarely make it to the $1 million box office.
Ang Lee's Crouching Tiger, Hidden Dragon won four Oscars and grossed $128 million at the box office.
On the other side of the globe, Chinese audiences, the world's largest film market, have to spend huge ticket prices on subtitles. At present, many theaters do not even schedule Chinese versions on imported films.
Taking the recently released "Out of Control Player" as an example, all cinemas and all shows in the author's business district are only available in English.
Humans usually read faster than they listen, so how many people inherently prefer subtitles over dubbing? For decades, the entertainment industry has been trying to find global markets by dubbing voices into movies, games, commercials, and TV shows. But it's a long and expensive process, and it's also accompanied by disjointed audience experiences and even piracy.
The rise of neural networks is quietly changing this traditional way of global cultural consumption. Machines want a "brave new world." All languages are "first-class citizens", there are no one-inch-high subtitles, not even human voiceovers, and you can watch "East Side Nightmare" as you would enjoy "The Twelve Hours of Chang'an". Eventually, the "Best International Film Award" could disappear and the Oscars simply become a Grand Prix without Borders.
01 Who's making the big cake?
The rise of commercialization of technology has always been inseparable from a fundamental premise - the market. Who needs it? Who made the demand? How impressive is this demand?
Translation must be due to the existence of a large market for non-English content demand. Netflix has been instrumental in creating a new global market for non-English-language content. Keep in mind that five to ten years ago, there was no demand for non-English content in the global market.
Founded for more than 20 years and experienced in hundreds of battles, Netflix can have today's market position, and it is indispensable to go overseas that year. The company began its overseas expansion in 2011, and the huge success of the 2013 home-made drama "House of Cards" not only made the company more convinced of the strategic importance of originality, but also recognized the limitations of subscriber growth in the United States. Netflix poured a lot of cash into content originality in the local language.
The self-made drama "Narcos", the British drama "The Crown", the German drama "Darkness", the Spanish "Banknote House" have achieved "blockbusters", the Portuguese drama "3%", the French drama "Marseille City", and the Japanese drama "Spark" have caused a sensation, all of which show that the audience in the overseas market has long matured, but the market lacks localization content suitable for them.
Currently, a large portion of Netflix's original content is non-English-language, with more than 60% of paying users coming from overseas. In its fourth quarter 2020 earnings report, the company announced that the vast majority of its recent subscriber revenue has come from abroad.
Netflix's business model, Global Subscription video-on-demand (SVOD), only really makes sense when it comes to spreading content across regions, and one of the biggest hurdles now is localization.
In the campaign to persuade audiences to overcome subtitle barriers to watch foreign-language films, Netflix has also done its best. In order to better distribute overseas, they added 28 language subtitles to the platform's content.
"We're seeing more and more subscribers choosing TV and movies that transcend borders and cultures." A Netflix spokesperson once told The New York Times.
The spokesperson also noted that subtitles don't stop good content from being loved and appreciated on Netflix. For example, Netflix's original series "Narcos" has both English and Spanish scenes, and the show has been renewed season after season.
Netflix original series "Narcos"
As the pie continues to grow and many competitors enter the game, Netflix ushered in more fierce competition Round 2. Major U.S. streamers are starting to expand their investments in content production, hoping to gain an edge in the new wave of SVOD subscriber competition. Among them, there are many traditional giants such as HBO Max and Disney Plus, as well as apple TV Plus, Amazon Prime, and other technology giants who are not bad at all.
Disney, for example, is shifting its core strategy from pay-TV, announcing its intention to shut down 100 international TV channels and migrate content to Disney Plus.
Amazon's budget for TV series, movies, and music on Prime has increased by more than 40 percent from 2019 to 2020.
However, Netflix is well aware of its first-mover advantage in non-English-speaking markets and foresees that overseas users will eventually become the core pillars of its business empire (like Google), and they choose to "surround the American mainland from overseas." In the words of Debra Chinn, Netflix's international dubbing director, "it's really creating a new audience."
This road is directly related to the future revenue space, and the dubbing work is therefore particularly important. For the past few years, they have been recruiting voice actors and producers to try to build an engineered production line.
On the one hand, it aims to improve the quality of the English version of international programming, attract English-speaking viewers who may be bored with low-quality dubbing, and improve the overall viewership of non-English series content by 371 million native English speakers (most of whom live in the United States) worldwide.
Dubbing content is an unfamiliar concept to AMERICAN or BRITISH TV viewers unless you watch Chinese martial arts films. Interestingly, market analysis firm Statista found that 59 percent of U.S. adults prefer to watch foreign-language movies dubbed in English than originals with subtitles.
In a 2019 interview, Netflix revealed that the dubbed version of the popular show was more popular. For example, 85 percent of U.S. viewers chose the dubbed version of Rain over the subtitles; 78 percent chose the dubbed version of the German thriller Dark; and 72 percent chose the dubbed version of Spain's Money Heist.
First Danish drama series "Miserable Rain"
Spanish drama "The House of Paper Money"
On the other hand, Netflix also wants to extend its English dubbing experience to more language dubbing. According to Ampere Analysis, 90% of the content Netflix broadcasts in most non-English-speaking regions is composed of foreign language content, and the dubbing work is important and arduous.
Take Europe, for example. Dubbed versions of shows have long been commonplace in Europe's television market, where shows often cross borders — a booming industry with its own awards ceremonies and voice actors themselves famous. This makes local consumers more willing to accept dubbing.
Especially in France, Spain and German-speaking countries, dubbing is the first condition for entering the market. In Germany, dubbing has also become a big industry, with top voice actors making $15,000 to $20,000 per film. If a major Distribution Company in the United States wants to successfully break into the German market, it must be dubbed.
Dubbing is also a very mature market in Italy. Today, 99% of Italian movies and TV shows are dubbed.
In recent years, Netflix's investment in dubbing has grown at an average annual rate of 25% to 35%, and the consumption of dubbing content has grown faster than investment, averaging more than 120% per year. The "financiers" increased their budgets, and the income of some large localized studios also soared.
Recently, ZOO Digital, a media localization provider with big streaming customers such as Netflix, Amazon, Apple, etc., received its best quarterly earnings report ever. In terms of market segmentation, localization contributes more than 51% to business revenue. Demand for the company's cloud-based dubbing services soared, with dubbing revenue up 50 percent to $9 million.
Another global localization studio, BTI Studios, accounted for just 3% of its revenue in 2010 and has since jumped to 61%, while subtitles earn less than 40%.
ZOO Digital believes that the streaming boom is good news for media localization vendors, as subtitles and dubbing play a key role in the ongoing audience battle between major streaming platforms. As content budgets increase, they find that buyers are increasingly looking to work with vendors who can provide end-to-end services.
Recently, BTI Studios merged with another localization company, IYUNO Media Group, to address "the rapidly growing demand for localized content by enabling global dubbing and subtitling capabilities," including the application of AI to respond to changes in the delivery cycle.
02 Dubbed Achilles Heel
Greg Peters, Netflix's chief product officer, attributed the increase in international film ratings entirely to better voice acting.
What is good dubbing? "The best compliment you can get as a voice actor is that no one realizes it's a dubbed sound," said Dietmar Wunder, a German who voiced Daniel Craig in the Bond series of films.
"Our job is to disappear behind the original so that the audience doesn't remember that."
However, most dubbing is still far away from "no trace". Scott Mann was stunned by a dubbed version of the film after directing the thriller Heist, starring Robert De Niro, in 2015.
The entire dialogue in the film has been modified to make the foreign language lines closer to De Niro's lip type. During filming, De Niro would spend weeks discussing a scene, asking what kind of cufflinks or socks the characters would wear. However, when the film was dubbed, his performance was tarnished.
"It's heartbreaking. We struggled with the script for months, arguing about which lines should be said and how to say them, only to see a foreign language version of it and the lines all changed. Everything feels terrible."
With existing technology and operating models, the results of Scott Mann's heartbreak are almost inevitable.
After the international distribution company receives the ADR (the technical processing method of bringing the actors back to re-dub the lines after the whole film is edited), the vocal cords and the final language vocal cords will be sent to different regions for translation and production.
In China, for example, a dubbing director looks at the original film first, and the translator translates the lines into Chinese text. Next, the porter conducts a preliminary check and adjusts the Chinese translation according to the length of the foreign text, the type of mouth, etc. Subsequently, the dubbing director, voice actor and translator watch the original film together, assign roles, and Chinese translation and re-pairing according to the actor's dubbing habits.
The best voice actors often modify the language to find the closest synchrony with the actor's mouth movements. For example, Hello, say "Hello." However, the English pronunciation of "llo" can be dragged, and in order to maintain the consistency of the lips, "good" must also be dragged, so there is a so-called "recitation cavity" of "hello".
Despite a lot of post-production work, there is still a mismatch between the actor's performance and the dubbing. Many Chinese audiences cannot appreciate this kind of "recitation".
The same is true in the field of video games. Many times, game publishers will only synchronize character lipography for English in their titles, and will do additional dubbing when localizing for other languages. If a sentence of dialogue is translated from English to French, the translated version may be much longer than the original sentence, usually by lengthening the animation to fill the extra audio time.
In addition to technical means, whether it is cost-effective is also an important cost factor restricting the quality of dubbing.
It takes an animator about 7 hours to animate a single character speaking for 1 minute in the game. Cyberpunk 2077, a large RPG with a lot of dialogue, would have to pay huge manpower and time costs if it supported lip syncing dubbed in 10 different languages.
In the United States, a two-hour movie can cost more than $150,000 in one language, blockbusters are usually translated into 12 languages, and TV series are often dubbed in 6 to 8 languages. The time and money required is not cost-effective.
Take the translation of "Conscience of the Industry" "Kung Fu Panda 3" as an example. It is said that Chinese dubbing took 8 months, and DreamWorks also set up a special Chinese creative team, including Chinese writers, directors, etc., who rewrote a version of the grounded dialogue in the way Chinese spoke. Chinese version re-spoken and emojied each character according to the Chinese lines.
Such time-consuming and labor-intensive operations can only stay at the individual level under the trend of many imported blockbusters being released to achieve quasi-synchronization (sometimes even early release).
As the post-production time becomes more and more tight, many times, the dubbing time of an imported animation film in China is only 3-5 days. It is said that a traffic star took only one day for "How to Train Your Dragon 3". Rough rushing can only perfectly destroy the original connotation of a beautiful movie.
With 10 60-minute episodes each, Netflix typically takes 16 weeks to dub. No matter how rich it is, the company also has a strategic focus.
In the Japanese market, almost every show has Japanese subtitles, and more than 40% are dubbed. Miyazaki and Studio Ghibli's wonderful works have been added with 28 subtitles and 20 dubbings.
In addition to Japan, the other four largest dubbing markets are in the European Union – France, Germany, Italy and Spain – where 60% of foreign content is dubbed. For Netflix, content dubbed in languages such as French, German, Italian and Spanish can also continue to be delivered to other markets such as Latin America, Africa and Canada.
But for example, in Russia, Turkey, where viewers make up only a fraction of Netflix's user base and Estonia has a population of only a million, it's hard to justify investing in their massive localization at the moment.
The most interesting is the Indian market. CEO Reed Hastings predicted that Netflix's next 100 million subscribers would come from India. But India speaks 450 languages, of which 23 are considered official. Since Netflix's subscription data in india is growing very rapidly, the difficulty of dubbing translation will have to be overcome sooner or later.
03 Deep Video Portraits Deep Face Swap
Where inefficient and intensive labor is gathered, there is a possibility of being revolutionized by new technologies.
The COVID-19 pandemic has further accelerated this process. Netflix had to postpone the local language versions of certain shows in 2020, with studios closing and voice actors being eliminated.
AI can help bridge those gaps, even if the studio ultimately decides to rely on real voice actors to make the final version. Startups with a keen sense of smell are entering with deep learning technology.
Like Disney, Scott Mann began working on Deepfake face changes. A basic assumption of Deepfake is that if you have enough 2D photos from multiple angles, you can actually create a complete, continuous 3D model.
You can train the model with a large number of 2D pictures of the star collected online, teach the AI to create a 3D face about the star, and replace the face of the target video by overlaying it.
However, in just a few seconds, most viewers can perceive that the video is fake. Because the training data is a static picture, when you paste the face of a celebrity on the face of a pornographic actor, it is easy to have eyebrow mismatch, unstable movements and other flaws.
In 2018, the Christian Theobalt team at the Max Planck Institute for Informatics demonstrated deep Vedio Portrait (DVP) at SIGGRAPH, which can reproduce near-realistic human visuals.
The system reconstructs the face using a series of face landmarks so that it can track and capture the subtle movements of sensitive areas (such as eyebrows, nose, mouth, etc.) in order to creatively "transfer" these subtle details to the target actor, so that his facial movements look as accurate and synchronized as the expression at the time of the new pronunciation. The researchers used a facial representation method to calculate the facial parameters for the two videos.
To make it more realistic, they also slightly modify these parameters for rendering. This is also where the Generated Adversarial Network (GAN) comes in. One neural network generates content, while the other rejects or approves each effort. The goal is to generate a fake image that is as good as the image in the target video frame to trick the discriminator network.
Just about two thousand frames — the equivalent of a minute of footage — is enough to train the network.
Unlike Deepfake, which uses 2D images for learning, DVP is trained in 3D facial movements of live actors. DVP is not a face swap, but a facial feature processing, like an AI-manipulated facial muscle puppet show.
DARPA official Matthew Turek has told foreign media that methods that can detect subtle clues to changes in current GAN works do not seem to be effective for DVP works. Because, it seems to perfectly match the entire facial and head movement between the source actor and the target actor.
In addition to the picky realism of the big screen, what impresses overworked visual effects artists and Hollywood producers on tight budgets is the cost performance. Production is fast and inexpensive, especially compared to the cost of a full remake of the film.
Digital face swapping is a common thing in the film and television industry, and the shooting of special effects shots has taken many years. In The Curious Affair of Benjamin Button, brad Pitt's face is replaced by a modified computer graphics version in almost every frame of the film.
Computer graphics editing of human faces has been widely used in today's feature films. A good example is The Wonders of Benjamin Button.
However, traditional special effects shots require a lot of post-production and consume huge capital costs, even for a few seconds of footage, requiring human and material investment, and often requires weeks of work by trained artists.
For example, stunt stuntmen will briefly look into the camera, and even for a brief moment, it takes a lot of post-production to make sure the people on screen look just like they're supposed to be.
Solving these problems often requires re-shooting, or a clever combination of computer mapping and compositing, which is never cheap. DVP shows how it's easier to create this visual effect in the future. One can easily edit the position of an actor's head and facial expressions to change the camera angle or subtly change the scene composition.
The researchers say that AI can not only produce highly realistic results, but it is also much faster than the manual processes used today.
Although the technology is still in the proof-of-concept stage, scott Mann decided to establish Flawless after discovering this research white paper, given its potentially disruptive impact on the visual entertainment industry.
As an AI company that flaunts the science and innovation of a new generation of filmmaking, the company's product introduction is indeed written in a similar way. The first (and only) AI product on the site is currently called TrueSync, which claims to be the world's first system to use AI to create perfect lip sync visualizations in multiple languages.
TrueSync has a so-called performance retention engine, which is a 3D model of the actor's head. The original video works involve details of the actor's unique performance style, such as frowning and the performance of the eyes hanging low, which can be learned and captured by the AI and preserved in its entirety. Although the actors' faces have been replaced, their original subtle performances will be preserved and carried over to the target faces to maintain the texture and immersion of the original film.
The technology automatically generates lip movements that correspond to the dubbed dialogue. Jack Nicholson's lip-syncing with Tom Cruise in The Wind is also almost perfectly synchronized with the French dubbing.
If the client needs Scarlett to say Chinese, they need to first record the Chinese into an actor's dialogue, like a traditional dubbing, and send those audios and pictures to Flawless. Based on this information, the company's system calculates Scarlett's unique mouth movements when she says Chinese, based on existing Scarlett models, to create a model of Scarlett's face Chinese. Finally, the company's digital effects art team was needed to polish it.
In Floless's view, the technology will be a more economical approach for directors who want to retain the charm of the original cast.
The 2020 Oscar-nominated Danish film Druk (also translated as The Alcohol Project, Drunken Tincture) was successful in China and was ready to be remade for an English-speaking audience with Leonardo DiCaprio as the lead actor. That means a $60 million remake cost.
If the male protagonist does not have enough appeal among English-speaking audiences, he can find a remake of a big name (for example, a Hollywood remake of "Infernal Affairs"). However, the male protagonist Max Mikkelson is not lowly famous in the English-speaking world, and the Danish drinking culture expressed in the film is no stranger to American audiences. If you use Formulaless technology, you can reach more English-speaking audiences at a lower cost (for example, 2% of the cost of remakes) while retaining the original flavor of the film.
04 Sound Art
Flawless's technology isn't as flawless as its name suggests, 100% flawless, but it's already pretty good. Companies like Disney are also working to improve the quality of deepfake technology so that it can quickly meet Hollywood's needs.
Although deep face-swapping technology has not yet achieved commercial success, voice AI has verified commercial success in the field of game video.
The highly anticipated sci-fi RPG Cyberpunk 2077 sold more than 13.7 million copies worldwide in its first month. It is said that the country with the largest number of pre-order players is China. This is not unrelated to the excellent performance of Chinese localization. When you play this game, you will find that the Mandarin-speaking characters are indeed like they are speaking Mandarin.
Voiceovers that respond to 10 languages are also seen as an important part of marketing. In the view of game maker CD Projekt Red, in the game, the most intuitive feeling for players is dubbing. By providing players with a localized experience, new retail markets can be expanded.
The technology of Jali Research, the technical support of the game, can detect the individual sounds that form each word and animate the character's face accordingly.
Just as Scott Mann was inspired by the ACC paper to start Floless, CD Projekt Red also found a technical muse on SIGGRAPH.
The game's localization is supported by a Canadian startup, Jali Research. They developed a set of tools that use AI to accurately map phonemes to the mouth shape of the target character, allowing the character to convincingly converse in different languages. The technology can be traced back to a study during the period of Dr. Pif Edwards, the company's chief technology officer.
Beginning his Ph.D. in computer science, Edwards wanted to focus on facial animation, but ultimately focused on speech, because "the results show that when people express themselves, they are almost always talking."
Not satisfied with the tools available at the time to handle speech and animation, he decided to build his own. In 2016, Pif Edwards presented a paper to SIGGRAPH discussing JALI, a model for expressing the synchronization of the mouth shape of pronunciation, and co-authored it also included an Oscar-winning animator and director.
Using the findings of psycholinguistics, they sought to figure out hundreds of variables that affect facial movement and created a network (VisemeNet) to predict the state of mouth shape in the target pronunciation. The so-called "JAW And Lip Integration", a combination of "jaw" and "lip", these two anatomical features explain most of the changes in visual speech, the paper says.
In the case of Cyberpunk 2077 dubbing, they combine machine learning and rules-based AI.
The first phase mainly uses machine learning. Suppose there is an audio file where someone says "Hello", where does "H" start and stop? Where are the "e", "l" and "o" sounds? They labeled this phoneme information into a specific language, and then used that data to train the machine, teaching it to recognize the sound someone actually makes when they speak.
Feeding a machine a conversation that has never been seen before, the system predicts the boundaries between sounds, as well as the length of each phoneme.
The second phase is mainly animation, and they use traditional rules-based AI. This is a simpler "if-then" system that simply performs the action being told based on a specific input and can be used to determine what mouth shape needs to be generated based on the sound emitted.
Contrary to most people's intuition, the shape of a particular letter or pronunciation of the mouth does not correspond directly to one to one. For example, the pronunciation of "en" often depends on the letters around it, rather than the pronounced mouth shape of the phoneme. If there is an "e" behind it, a "n" or "noo" sound may be emitted at the end.
The so-called rules refer to some rules that are followed when making facial expressions of the same sound, no matter which Chinese is spoken. This means that you need to create language models for different languages, but you don't need to create separate animation models for different languages, and the animation components of all languages are the same, which can save a lot of time and cost.
In fact, in addition to realistic effects, studios can scale facial animation at a relatively low cost compared to more expensive traditional methods such as keyframing and motion capture animation.
Jali Research's localization is not just lip-sync, but even takes into account facial expressions, which is convincing.
Also radically dedicated to the art of speech is Deepdub, an Israeli company incubated by Silicon Valley YC. If Flawless's technique retains voice actors for the time being, Deepdub simply uses the actors' original voices to interpret the local language, even though he never says the lines.
In one video it has shared internally, Morgan Freeman's English-speaking video can be switched to dubbing mode, in which Freeman himself speaks Spanish — not through a human voice actor, but with his own husky voice.
Developing this system requires a combination of different technologies, including deep learning, signal processing, unique actor voice modeling, neural style transfer, and NLP. Among them, deep learning learns the characteristics of sound in the audio track data, including pitch, timbre, speed, spacing, and intonation, and records them for use in new lines.
The company is also trying to match lip sync or lip movements to further enhance the audiovisual effect.
The founders are two movie enthusiasts who want customers to easily localize multilingual content by pressing the select button, leaving everything behind it to AI.
Although the company has yet to prove itself with a publicly available product, the potential behind its publicity is clear. More recently, Deepdub also successfully invited former HBO Max Chief Content Officer Kevin Reilly to join the company's advisory board to help bring its AI dubbing technology to Hollywood.
If Deepdub's AI is deployed at scale, it will disrupt the traditional rules of the game for studios, content creators, and producers. Kevin Reilly believes that this is almost inevitable. He even predicted that he would not return to the tv executive ranks.
For example, it could lead to job losses for employees in the dubbing industry, and traditional ADR might not be needed, because synthetic voices can replace actors, and post-production does not have to return to the studio to re-record certain lines.
Now, a TV series can take 14 to 16 weeks to dub, while Deepdub's technology takes 4 to 6 weeks and is planned to be further compressed to less than two weeks, which costs much less than manual dubbing.
These emerging technologies can also effectively prevent piracy by significantly speeding up dubbing cycles and shortening the window between local and other regional releases. This has been proven in the game market. As CD Projekt Red put more emphasis on translation and localization, more and more Polish players are willing to pay for genuine games.
At present, it is very important for these AI companies to partner with major Hollywood companies to profit from technology.
Everybody who saw the show said, "Wow, I can speak French." When Floless showed the actors a performance after being dubbed by AI, they were stunned. Flawless is in contact with film companies and is also working with one of the largest streaming platforms.
Deepdub also began to get serious. They struck a deal to use their AI TTS technology to pair the film "Every Time I Die" with Latin American Spanish and Brazilian Portuguese, which is currently being broadcast to American audiences on UK Netflix.
05 Special effects and forgeries
When Leonedo DiCaprio and his girlfriend were standing at the bow of the ship, the camera lens pulled back two or three miles away, and it was done seamlessly.
The film industry uses special effects, but no one realizes that similar technology is inherently a forgery.
For decades, the quest for the magic of technology (efforts beyond the effects of previous films) seemed to be the supreme goal. This manipulation of the real has led to a new kind of reality— hyper-reality. Whether it's computer synthesis, digital image processing, simulated 3D animation, or artificial intelligence, filmmakers' quest for hyperrealism has reached an irreversible level.
Now we can easily generate a very realistic human face. The next question is, how should we deal with these generated faces? Facial treatment can cause controversy in the film industry. How do some actors react when they see their performances change? How to make this technology legal and ethical? How do you view a reality that is dissolving?
To be sure, whether it's anonymous Redditor, big players like Adobe, or academia itself, all channels are making progress in advanced audio and visual processing, and no one promises to put the brakes on. The research community is very good at sharing technology openly and widely, and the rapid democratization of power is almost inevitable. AI effects are only going to get better, cheaper, and faster.
Film directors and producers are also showing increasing interest in these technologies. This situation is similar to the 2000s when using realistic computer graphics. Thanks to AI, "suddenly everybody wants to do something."
We are in an era where anyone can make a believable video, and anyone can claim that the video is fake. It is inevitable that society will have to adapt to this change, albeit painfully.