Produced by | Tiger Sniff Youth Culture Group
Author | Mu Zitong
Editing, Cartography丨Slag County
This article was first published on the Tiger Sniff Youth Content public account "That NG" (ID: huxiu4youth). Here, we present the faces, stories and attitudes of today's young people.
"I hid 200 million yuan worth of gold in a certain sea area, do you want my treasure?
Zhang Meilan, the richest woman in Vietnam, who was sentenced to death a few days ago, made "One Piece" come to life with one sentence. According to a video of the trial that went viral, she admitted to hiding a huge amount of property at sea.
As soon as these words came out, it was like a shot of adrenaline poked into the arteries, and the Vietnamese people who were drowsy in the spring suddenly became energetic, and even Chinese coastal netizens were ready to worship Mazu and set sail.
Who is Zhang Meilan?
Chinese businesswoman, real estate tycoon, billionaire. He was recently sentenced to death and fined VND 673 trillion for being involved in the largest financial fraud case in Vietnam's history.
Even in terms of the extremely inflated currency denomination in the local area, 673 trillion yuan is a huge amount of money that exceeds the perceived limit, equivalent to about 192.4 billion yuan.
In the trial video leaked on the Internet, the adjudication committee asked Zhang Meilan about the whereabouts of the huge amount of money that disappeared, and Zhang Meilan replied calmly and slightly provocatively, all the money is at sea:
"Does anyone know where the defendant hid the money?"
"No one knows yet. ”
"Therefore, the defendant must honestly inform the adjudication committee in order to receive leniency. ”
"Yes yes, at sea, yes, that's right, there are six hundred and seventy-three trillion dong......"
The "200 million yuan" rumored on the Chinese Internet is still too conservative, and according to the source video, if you find this treasure, it will really be "rich and rival the country".
The death penalty, the treasure, the sea, Zhang Meilan's big secret treasure is like the beginning of "One Piece".
One Piece Roger, who possessed endless treasures, publicly shouted before his execution: "Do you want my treasure? ”
The world of "One Piece" thus opened the "era of pirates" with thousands of sails.
As soon as the trial video came out, Zhang Meilan seemed to have become a contemporary Roger. The entry "Looking for Zhang Meilan's treasure" rushed to the hot search in Vietnam, and when you click on TikTok at this time, you will feel that all Vietnamese friends have washed into the sea.
Some people sail away with their own fishing boats, some people wear diving equipment and form a group to touch gold, and the technical stream boss analyzes and investigates the location of the gold that may be hidden, and the experience stream brother asks the fish for direction......
If you don't have the money to go to sea, batch a basket of dragon fruit, resell "Devil Fruits", and if you have an equipment base, buckle a straw hat and start cosing Luffy.
Even Google Maps has many addresses called "Zhang Meilan's Treasure" out of thin air, but if you follow the trail, there will only be a small restaurant in the place, and the owner of the shop will smile narrowly.
No, wait, how can this matter become more and more interesting and smelly?
There's nothing wrong with that, as you feel about the sense of disobedience. The Zhang Meilan Secret Treasure incident is an outright fake news.
Zhang Meilan's defense lawyer, trial participants and other people have successively proved to the media that on the day of the trial, Zhang Meilan did not say anything about "hidden treasures at sea".
The trial video that went viral on the Internet was completely faked, and judging by its smoothness, it is likely to be with the help of AI deepfake technology.
The outrageous 192.4 billion treasure has never existed at all. But ordinary life is too boring, some people know it's fake, some people half-believe it, Zhang Meilan's treasure is just half-pushed and half-pushed by the people who eat melons, and it quickly became popular. When it was transmitted to China, it evolved a local adaptation version of "200 million gold".
AI deep disinformation has deceived us again and again, so easily.
Proficient in AI mapping, each hand is a gold brick
A year ago, when AI technology exploded, we raised a lot of dystopian anxieties about its application prospects, such as AI stealing people's jobs and AI controlling humans. A year later, all of them are still in the distant stage of prophecy, and only the AI deep pseudo-information anxiety is quickly falling into reality.
When AI technology greatly reduces the cost of information fraud, everyone can easily generate realistic news pictures and videos, and AI deep fake information is proliferating rapidly, and "deep fake" is no longer a distant topic in the future imagination.
It's rapidly invading reality, changing reality, becoming reality, and most of us are far from being prepared.
In the era of AI hypocrisy, "seeing is believing" no longer exists.
Because the cost of counterfeiting is infinitely reduced.
A few months ago, in a small town called Almendralejo, more than 20 junior high school girls suddenly discovered that their classmates were circulating "nude photos" of them. The naked unfamiliar body and my face are undoubtedly AI deep fake photos.
The suspects were their classmates, several equally young, ordinary junior high school boys.
A few adolescent hairy boys can also create false information that is enough to mess with reality, pointing to the most terrifying point in the era of AI deep fake: "fake" has been reduced from a craft that needs to be learned to a tool that everyone can use at will.
You don't need to learn any technology, as long as you have a "NUCA camera", you can "undress anyone with one click".
Aim at the subject, click, and after ten seconds, the NUCA camera will spit out an exclusive "nude photo". This is a special installation by artists Mathias Vef and Benedikt Groß.
The imaging principle is very simple, after the camera takes a photo, connect to the Stable Diffusion, and ask the AI to generate a corresponding "nude photo" with reference to the photo. The resulting image is certainly not the body of the person being photographed, but at first glance it is quite convincing.
In doing so, the artist hopes to remind people of how easy it is to generate a fake pornographic image that invades privacy.
The low cost of counterfeiting comes at a high cost. It takes years or even decades to maintain a healthy relationship, and it only takes a few seconds of AI deep fake video to destroy it.
In a Netflix love variety show called "Deep Fake Love", a close couple who have been in love for 9 years did not bear the test of AI deep fake.
The program team dismantled the couple and threw them into the villa surrounded by hot singles and the opposite sex, and then showed them a video of each other's lives every once in a while.
The video content is usually very "excessive", sometimes the partner is flirting with the single opposite sex, sometimes saying bad things about the other person, and sometimes even rolling the sheets and kissing the single opposite sex. There are real and fake, and the fake is generated by AI deep fake.
Spanish couple Manuel and AIDA are ready to get married after 9 years of long-distance love, but after only two days of separation, Manuel sees his fiancée fall in front of a handsome bachelor: the two look at each other in the pool for a few seconds, and then kiss together.
Manuel was shocked and couldn't believe her eyes: "What? Who is this girl? Who is she? She is making out with a man! They are making out!"
Although the host soon told him that the video might be AI, Manuel couldn't let go of it for a long time - the video was so realistic.
The rough hair and water droplets glide through the slightly textured skin in the sun, and the muscles rise and fall softly with the movements, and there is no trace of AI in the careful production.
The reality is that AIDA awkwardly smiles and awkwardly avoids the attack of the handsome bachelor, and in the deep fake video, her movements are tampered with to cater to the man's hot kiss.
"Deep fakes have been around for a long time, and the big difference now is that they're getting better and better. ”
Think about it, if we see such a video not on a TV show, but in real life, can we still maintain trust in our partners, friends, and partners?
In real life, no one will reveal the correct answer at the end. And the reason why deep hypocrisy is terrible is precisely because once the seeds of doubt are planted, they will suck and uneasy to take root and sprout on their own.
AI hypocrisy is not only hoaxes and sexual insults, but can even hinder the country's political activities.
Back in 2018, a team of researchers at the University of Washington used a video of Obama's speech to predict the horror of the emergence of deep technology in the political realm.
In the video, Obama, who is synthesized by deep pseudo technology, says vividly and naturally:
"We are entering an era in which our enemies can make anyone say anything at any point in time – even if they never say those words. Like, they can ask me to say something like this. ”
At the right time and in the right place, AI Deep Pseudo only needs to gently apply a little thrust to achieve the effect of four or two thousand pounds.
During last year's elections in Slovakia, the leader of the Progressive Party, Michal Šimečka, suffered greatly. Forty-eight hours before the vote, a recording of him contacting reporters to discuss ballot manipulation popped up on social media platforms.
In the audio, he mocks his voters in a bland voice and makes a low-level joke about child pornography.
Daniel Milo, head of the Slovak Interior Ministry's Department for Combating Disinformation, said the recording couldn't have been faker: "You can definitely tell it's not a real person – but the recording went viral." ”
Clearly, this is a careful plan with insight into human nature and rules. Soon, a number of fact-checking departments confirmed that the recording was a deep fabrication. But Slovakia's election rules mean that the media and politicians should remain silent during the critical period leading up to the election, making it difficult to spread rumours widely.
AI deep fake technology is far from perfect, but it is enough to convince those who want to be convinced.
The motivation to "believe" is always accumulated in the shadows, and it only takes a small support to roll it on its own.
In the same year that Obama's fake video appeared, an attempted coup d'état in the African country of Gabon was a perfect proof of this.
Then-President Ali Bongo Ondimba was absent from public life at home for a long time due to medical treatment abroad, and the government was secretive about his condition. Some began to question his physical condition and even suspect that he had died.
Amid the clouds, the president delivered a video New Year's address. The original intention was to refute the rumors: Look, I'm still alive!
It backfired, and he looked weird, stiff, and full of disobedience in the video. So the opposition immediately concluded that the video was faked and quickly staged a coup d'état - the president may have died, and the government is secret!
"Those who stubbornly cling to power, once again use the president in order to save a sick man who has lost his physical and mental capabilities. ”
Coup leader Obiang said with such conviction.
Unfortunately, the video is not fake, and the president is still alive and in good spirits. The coup d'état was quickly thwarted, and the plotters were either captured or killed. Fact-checking experts after the incident pointed out that the stiff and strange look may have been caused by the president's Botox injections and excessive makeup.
The regime subversion triggered by this deep hypocrisy was unsuccessful, but no one can say what will happen next time.
We have entered an era of deep symbiosis with AI deep pseudo-information. Any ordinary person around you can easily create deep disinformation. Their intentions may be malicious, or they may be just a joke, but once false information appears, its circulation is no longer controllable.
When an atmosphere of suspicion created by AI is in the air, public figures can even use it as a way to evade responsibility. Imagine if Watergate had happened today, Nixon would have been able to claim that the key audio evidence was falsified by AI.
Communication scholar Robert Chesney has figuratively called it "the liar's dividend."
The unprecedented deep pseudo-anxiety of AI has made that once distant problem suddenly become real and urgent:
Do we have a way to control the deep fakeness of AI, and how can ordinary people survive in such an era where it is difficult to distinguish between truth and falsehood?
Technically, completely controlling AI counterfeiting is like trying to control pirated products, which is very difficult.
The most striving is the major social platforms that provide distribution channels. Because of the European Union's Digital Services Act (DSA), companies are legally responsible for harmful content, false information, etc., and if found to be in violation, they will face fines of up to 6% of their global turnover.
In order not to be fined money, Meta, X, TikTok, YouTube have all put on a fairly positive posture.
They are working with universities to develop AI disinformation identification tools, set up countermeasure operation centers, and put mandatory labels to identify potentially AI-generated content. Deepfake videos were deleted as soon as they were discovered, and Reddit even shut down deepfake topics with more than 90,000 participants.
But the seemingly reliable platform rules are like Du Fu's roof, leaking everywhere.
First of all, accurate detection is an impossible task. As Lusia Verdoliva said, "You can't make a universal detector – it's too hard." ”
She is one of the developers of FaceForensics++, a deep fake video recognition software that can detect the claws of most commonly used fake software, but it is not 100% probable.
Like a cat-and-mouse battle between a treasure connoisseur and a counterfeit dealer, every time a technologist finds a secret to counterfeiting, the deep counterfeiter quickly updates and iterates.
At the same time, when killing "deep pseudo-cancer cells", the platform must avoid accidentally damaging the real information, so the censorship is always "raised high".
In other words, in the days ahead, we are already destined to coexist with a sea of deep disinformation.
So will ordinary people lose their judgment of truth and falsehood because of deep falsehood? A small test, come and test your quality.
The following 6 pictures are either real photos or AI deep fake, which ones do you think are true?
The answer is announced: false true true false false.
Image from the online test "Deepfake, Can You Spot Them?" produced by the Kellogg School of Management at Northwestern University. The researchers tried to collect real human data to analyze whether humans can distinguish AI deep fake images and the basis for judgment.
Another European study found that 58% of 16-64-year-olds reported that they regularly interact with AI information, but only one in five expressed confidence in recognizing AI-generated content.
One in four other respondents said they believe they are regularly exposed to deep AI disinformation.
The researchers point out that the actual numbers may be worse than the studies suggest, because many people are exposed to AI information without realizing that it comes from AI.
How to distinguish the authenticity of information in the era of AI seems to be an emerging discipline.
When the media shifted from the mass media of radio and television to the online world, people began to talk about "media literacy in the Internet age": how to use online media correctly and identify online information.
In the era of AI, another media revolution, it is time to talk about "media literacy in the age of AI".
MIT has pioneered the online open course "Media Literacy in the Age of Deep Fake", which teaches people how to learn how to identify deep fake information.
This can be done with tools provided by governments and public institutions, such as Sensity's "deepfake detection" platform, the U.S. Defense Advanced Research Projects Agency's (DARPA) MdeiFor program, Adobe's Content Authenticity Program, TinEye's image recognition software, and Amnesty International's YouTube Data Viewer.
Or to put it more simply, look for credible traditional media and look at the judgment of fact-checkers and media professionals.
"Traditional media do sometimes spread misinformation, but they are still the most reliable source of information. ”
It's like the mantra of crossing the street, "one look, two slow, three pass", before clicking to share, doubt for one more second.
Somewhat ironically, in this age of conspiracy theories, the most scarce quality is "skepticism".
Multiple surveys have shown that "critical thinking" is considered the most scarce media literacy skill today.
Audiovisual Observatory Europe's statistics on 547 media literacy programmes in EU member states found that critical thinking was the most frequently cited skill. Another survey noted that 95% believe critical thinking is essential in today's world, but the majority of them (85%) believe that the public lacks these skills.
While we are suspicious of everything, we are always gullible about content that fits our stereotypical values – if a video shows Trump preparing to recruit prostitutes in public to express his dissatisfaction with feminism, then there must be many who believe it, after all, Trump can be so outrageous in our hearts.
It is impossible to tell whether it is true or half-pushed, and deep false information can always be widely disseminated under the mentality of "watching the excitement and not being afraid of big things".
At this point, an ordinary citizen with a critical mind becomes invaluable: they will play the role of a flame retardant in the chain of rumors.
What is really scary is never the AI deep fake, but the viral spread of the deep fake.
There is nothing new under the sun, and if you want to say that media innovation brings trust anxiety, in fact, it has happened in every media iteration.
When photoshop appeared, there was a P picture fake.
When radio first appeared, people believed the alien invasion stories on the radio.
The crisis of information trust brought about by these new technologies soon became commonplace and no longer caused anxiety.
Nowadays, whenever there are "nude photos" of celebrities, many people will ask in their hearts, "Is it P"?
As the public has seen more and more AI deep false information, every time there is too abnormal information in the future, the same sentence will be put forward to all reactions: "Is this AI?"
AI deep fake technology is indeed cause for anxiety, but there is no need to worry about it.
Just like the natural ecosystem, human society is constantly adapting to new changes in self-regulation, and the balance that is broken by new technologies in a short period of time will soon be restored to stability under the self-regulation of society.
In the final analysis, the problem of AI deep fakeness is not in AI, nor is it in deep fakeness. Like every previous counterfeiting technology, AI Deep Fake just provides a basis for choice. Technology is not guilty, the problem is still a human problem.
Why are human beings so keen to spread rumors? Why have rumors never stopped for thousands of years?
Because information is never just information, but also values, prejudices, and gestures.
Receiving information is passive and objective, disseminating information is a subjective behavior, and choosing which of the massive information to spread and share often depends only on personal likes and dislikes:
In a word, what you share is you.
People who are changing and want to change the world are all in the Tiger Sniff APP