Ancient "surveillance footage" has gone viral, the leaked "real shot" of Wu Song slaying the tiger is circulating widely online, and these AI-generated playful stunts have left netizens completely confused.
Recently, I came across a trending search about AI-generated meme content that must not mislead the public, and after reading the related content, I felt the world is becoming more and more surreal.
It started with a video titled "Surveillance Footage from the Hongmen Banquet Site in 206 BC Leaked". When you click in, you see a black-and-white picture with a time code jumping in the upper right corner. The camera is aimed at the banquet hall, where Xiang Yu and Liu Bang are sitting opposite each other drinking, Xiang Zhuang performs a sword dance and Fan Kuai bursts into the tent. Every movement looks incredibly natural.
Image source: Screenshot from Weibo
There were surveillance cameras in ancient times? I thought it was just the creator's creative spoof at first, but the more I scrolled, the more absurd the content got. There are "real surveillance footage" of the Xuanwu Gate Incident, on-site recordings of Wu Song fighting the tiger, and even the daily life of Yang Guifei in the Huaqing Pool is shot clearly.
It has to be said that the threshold for AI creation now is unbelievably low. In the past, to make a spoof work, you had to master video editing, photo editing, and search for materials for a long time to produce one piece of work. Now, by just inputting a sentence, AI can immediately generate a video for you in whatever style you want.
Various emerging AI meme creation trends
The most popular content right now is the ancient surveillance series. Creators use AI to generate footage of historical events, then add the look of surveillance cameras: black-and-white picture quality with some grainy noise, perfectly replicating the texture of old surveillance videos.
The earliest viral Hongmen Banquet video even vividly restores the small details from Fan Zeng raising his jade ring to Xiang Bo standing up to block the sword, with the caption "The archaeological team accidentally found the surveillance hard drive of the Qin Dynasty, and the whole internet was shocked after it was repaired". Relying on the gimmick of pseudo-documentary and exquisite picture details, this video quickly gained millions of views, driving the whole creative track of ancient surveillance content. Many creators followed suit, and later multi-angle surveillance footage of Qin Shi Huang's enthronement ceremony and Wu Song fighting the tiger at Jingyang Hill emerged, focusing on full documentary of ancient life.
There are also many people playing with memes in the comment section: No wonder the history books have such detailed records, it turns out there were videos. Everyone knows clearly that these are just spoofs, and they have a lot of fun with it.
Image source: Screenshot from Douyin
If the ancient surveillance series is still creative meme making, then the AI secondary creation of Jensen Huang has pushed the sense of contrast to the extreme and became a smash hit beyond the tech circle.
As most people know, Jensen Huang publicly admitted to being unfaithful to his wife, and the scandal with industry executive Huang Le caused a huge stir. It was originally a business gossip, but netizens used AI to turn it directly into a cheesy romance drama. Various secondary creations emerged in endlessly, paired with old web novel lines like "He controls the global computing power, but can't control his own heart", directly turning the tech giant into a cheesy romance male lead.
What makes netizens feel a mix of amusement and frustration most is the AI face-swapped version of "Trouble Maker". Netizens synthesized the faces of Jensen Huang and Huang Le into the classic pas de deux clip. On one side is the leather-jacketed giant standing at the top of global computing power, on the other side is the ambiguous stage dance music, the sense of incongruity is full, and it has gone viral on all major platforms in just a few days.
When I came across this content, I couldn't help but sigh that Jensen Huang probably never imagined that after working hard all his life to become the godfather of AI, he would end up being spoofed by AI.
Now think about it, AI meme creation has long penetrated every corner of the Internet.
Previously, many classic old dramas were modified in absurd ways. In *Journey to the West*, Tang Seng uses navigation to go on his pilgrimage for Buddhist scriptures, Sun Wukong uses a drone to scout the road, and Zhu Bajie uses a mobile app to order takeout; the concubines in *Empresses in the Palace* use smart phones and send messages to attack each other in the palace fight group. This kind of absurd modified content relies on the built-in audience base of classic IPs, spreads very fast, and many people who have never watched the original drama even remember the AI-adapted version first.
Image source: Screenshot from Weibo
In addition to visual secondary creation, the forms of AI meme making are also constantly innovating. For example, adding AI-generated funny dubbing to classic old films, making ancient characters speak authentic internet buzzwords and dialect jokes, creating a strong sense of contrast; there are also AI-generated absurd quotes from various celebrities paired with pictures to make short videos, many people can't help reposting them when they see them. Some creators even use AI to generate a large number of funny short stories in batches, with the copywriting, pictures and dubbing all completed by AI, and they can produce dozens of videos a day. Now some AI meme content can generate anything as long as it can attract attention.
If everyone knows that the content is fake, that's of course no problem. But the problem is that a lot of AI-generated content now, as it spreads, is gradually taken as real by people.
The boundary between true and false is getting increasingly blurred
Obviously these are meme content, why are they so easy to mislead the public? I think there are three reasons as follows.
First, AI generation technology is advancing too fast. A couple of years ago, AI-generated videos were full of flaws and could be seen as fake at a glance. Now, whether it is text-to-video generation or AI face swapping, human movements and subtle facial expression changes can be made very natural.
Second, the creators deliberately mislead the public. Many creators intentionally use misleading copy, such as "Precious surveillance footage leaked" and "Exclusive breaking news just unearthed". Even many marketing accounts, when reposting the content, deliberately do not mark that it is AI-generated in order to attract traffic.
The more novel and contrasting the content is, the higher the completion rate and repost rate will be, so AI-generated attention-grabbing videos are more likely to be pushed to more people. Some videos with all traces of AI marking removed are usually not processed until they have spread widely. By then, many people have already taken the content as real based on their first impression.
Finally, most people watch short videos to relax, and do not watch them with the mentality of distinguishing between true and false. When they come across an interesting video, their first reaction is that it is fun, then they like and repost it casually, without thinking about whether it is real. Anyway, it's just a meme, it doesn't matter if it's fake, as long as you are happy.
Image source: Screenshot from Weibo
This mindset itself is not wrong, but as more and more people spread the content, the fake will gradually seem real. For groups unfamiliar with AI such as the elderly and children, they have no idea that AI can make such realistic videos now, and are very vulnerable to being misled.
What is more worthy of vigilance is that police in many regions have previously reported cases of AI face-swapping fraud. Scammers use AI face swapping to impersonate the relatives and friends of the parties concerned, and defraud hundreds of thousands of yuan in just a few minutes. Even public figures are not spared. Some time ago, actor Wu Yue posted a late-night statement angrily denouncing that someone used AI technology to impersonate her to do livestreaming marketing, and many audiences could not distinguish the authenticity and were easily misled to place orders.
Similar situations are no longer rare. Many veteran actors and public figures have had their faces swapped into livestreaming marketing rooms via AI, without their knowledge, but they are forced to take the blame for counterfeit products. Different from ordinary creative meme making, this kind of counterfeiting is aimed at making profits, with extremely low counterfeiting cost. With only a few pieces of public video materials, a whole set of realistic livestream footage can be generated.
The trouble is that such livestreaming rooms usually change accounts and run away after operating for a few days. After consumers are cheated, it is very difficult for them to find the behind-the-scenes merchants, nor can they hold the impersonated artists accountable, so the threshold for rights protection is very high. If people keep using AI to generate fake videos to spread rumors and stir up trouble in accidents and social hot events, the consequences will be unimaginable.
Having said that, I think a lot of current AI secondary creation content has crossed the line. Many AI face swaps are done without the consent of the person concerned, directly putting their faces on other people's bodies. Whether it is a public figure like Jensen Huang or an ordinary passerby, this actually constitutes an infringement of portrait right and privacy.
If someone with malicious intent puts the faces of their classmates or colleagues on weird videos and posts them online at will, it will cause great negative impact on the people concerned.
Final Thoughts
In the past, only professional teams could produce video special effects, but now ordinary people can generate creative content with just one sentence. From the perspective of my official account Xiaolei Bilibili (ID: xiaoleibbb), this is a good thing in itself.
But all of this cannot be based on misleading and deception. Creators need to be more self-disciplined, and platforms should do a good job in content screening, and meme making cannot be done without any bottom line.
In the view of Xiaolei Bilibili (ID: xiaoleibbb), technology itself is innocent, the key lies in how people use it. The same is true for AI: it can help us create content and improve efficiency, but it can also be used to spread rumors and commit fraud. In the future, AI technology will become more and more advanced, and the generated content will become more and more realistic. When fake content floods the internet, we ourselves will suffer in the end.
Entertainment is entertainment, don't let AI confuse the truth.
Finally, I want to ask everyone: what do you think of the incidents where AI meme making gets out of control and confuses right and wrong? Do you usually like watching AI secondary creation videos? Let's chat together in the comment section.