AI-Generated “Authentic Fakes”
The Light and Shadow of Deepfakes, Explained.
Even if you’re not familiar with AI, no worries! By reading this article, you’ll fully understand what the hot topic “deepfake” technology is and how it affects our lives.
What you’ll learn in this article💡
You hear the word “deepfake” on the news but aren’t sure what it means? This article is for you! By the time you finish reading, you’ll clearly understand the following:
- How to create fake videos that are indistinguishable from real ones
- The uses of “good deepfakes” that are useful in movies and medicine
- The scary tricks of “bad deepfakes” that can easily deceive you
- How we can protect ourselves starting now
1. What are Deepfakes? ~ AI’s Magic Mirror ~
Have you recently seen incredibly realistic celebrity videos on social media? Perhaps they were “deepfakes.” In short, deepfakes are “AI-generated, authentic-looking fake videos and audio.” It’s like a magic mirror, capable of swapping one person’s face onto another or making someone say things they never did.
At its core is an AI technology called deep learning. Specifically, a mechanism called GANs (Generative Adversarial Networks) is often used. It might sound complicated, but the mechanism is surprisingly simple!
Dr. AI
Think of GANs like a competition between a “counterfeiter” and a “detective trying to catch them.”
You
A counterfeiter and a detective? Sounds interesting!
Dr. AI
The counterfeiter (Generative AI) tries to fool the detective (Discriminative AI) by making more and more realistic fake bills (fake videos). The detective works hard to expose the fakes. By repeating this competition millions of times, the counterfeiter eventually becomes capable of producing fakes so perfect that even the detective can’t distinguish them from the real thing.
You
I see! So, AIs compete with each other and get more and more realistic!
2. Amazing Uses for Deepfakes! The Bright Side ✨
The word “fake” often gives a negative impression, but deepfakes also have the potential to enrich our lives. You could call them “good deepfakes.” Let’s look at some specific examples!
Entertainment
In the Star Wars series, a young Luke Skywalker can appear, or foreign films can be dubbed into English with the original actor’s voice. It’s even becoming possible to bring deceased actors back to the screen, expanding the range of expression infinitely.
Healthcare and Welfare
People who have lost their voice due to illness can regain it from past audio data and talk to their families. Such a dream-like scenario is starting to become a reality. It’s also being used for realistic surgical simulations in medical training.
Education
Historical figures, like Einstein, could teach us physics as if they were alive today. Such interactive learning experiences are becoming possible, which might make studying even more enjoyable.
Advertising and Art
Even without hiring models or talents, AI can generate attractive advertisements using fictional characters. This can reduce production costs while allowing for diverse expressions in art and advertising.
3. Don’t Get Fooled! The Scary Traps of Deepfakes 👿
Just as there is light, there is shadow. Deepfakes also have very dangerous aspects. In particular, fraud and the spread of misinformation through “impersonation” are becoming serious social problems.
Case.1 Financial Fraud – $25 Million Victimization by “Voice” Alone
This actually happened. In Hong Kong, a company’s finance employee participated in a video conference. The British head office’s CFO and colleagues were supposedly there. However, it turned out that everyone except the finance employee was a deepfake!
Lured by Fake Email
A fake email, seemingly from their superior, instructed them to join an urgent video conference.
An Authentic-Looking Fake Meeting
In the conference, the superior and colleagues looked and sounded just like the real ones. The employee didn’t realize they were all deepfakes.
Top-Secret Transfer Instructions
The fake superior instructed them to transfer funds to multiple accounts, calling it a “top-secret transaction.”
Massive Loss Incurred
The trusting employee transferred a total of approximately $25 million. Later, upon checking with the real head office, they discovered it was a scam.
In incidents like this, when voices and faces are so indistinguishable from the real thing, it becomes incredibly difficult to even suspect anything.
Case.2 Misinformation and Election Interference
Deepfakes also carry the risk of fundamentally undermining societal trust. For example, what would happen if a fake video of a particular candidate making extremist remarks they never uttered were to spread during an election period? Voters would become unsure whom to trust, and fair elections could be compromised. This could pose a threat to democracy itself.
Case.3 Fake Pornography – Your Face Used for Crime
And among the malicious uses of deepfakes, the most serious and demeaning to individual dignity is “fake pornography.” This is an extremely malicious act where a person’s facial photos are used without permission to synthesize obscene videos and images.
You might think, “Only celebrities are targeted, right?” But that’s a big misconception. Now, anyone can become a victim with just one facial photo posted on social media. And once it’s spread online, it’s nearly impossible to completely delete, continuously tormenting victims as a “digital tattoo.”
⚖️This is absolutely not “play.” This is a “crime.”
The creation and dissemination of such fake pornography are unforgivable human rights violations that deeply wound the victim’s heart.
In the United States, laws are being developed to prosecute such acts.
Many states have laws in place, or are developing them, to address the creation and sharing of non-consensual deepfake pornography, often falling under categories like revenge porn or misuse of personal likeness. Never forget that actions taken lightly could ruin your entire life.
4. What Should We Do? 🛡️ Tips for Surviving the Deepfake Era
With such sophisticated technology, you might feel anxious about how to detect it. There’s no perfect countermeasure, but individuals can take steps, and society as a whole is starting to implement initiatives.
① What Individuals Can Do: Make Information “Health Checks” a Habit
If something seems off, don’t immediately believe it. That pause is crucial. Try checking the following points as if you’re giving information a health check.
🤔 Deepfake Quiz 🤔
This video seems suspicious. What should you check?
※Click the buttons to reveal the answers!
Above all, always be aware of “where did this information come from?”
Make it a habit to cross-reference information with multiple sources, such as official websites of public institutions or reputable major media outlets.
② Societal Efforts: Countering with Technology and Rules
It’s not just about us as individuals; society as a whole is also making progress in countermeasures. There are broadly two approaches: “technology” and “laws.”
| Approach | Specific Content |
|---|---|
| 💻 Technological Measures | The development of “detection technology” using AI to identify deepfakes is progressing. Also, systems like C2PA are emerging, which record “when, where, by whom, and how” a photo or video was created, acting like a “digital birth certificate” to prove its authenticity. |
| ⚖️ Legal Measures | Around the world, legal frameworks are advancing to address defamation and fraud using deepfakes. Concerns about electoral interference, in particular, are accelerating efforts to prohibit the creation and dissemination of malicious misinformation. In the United States, several states have already made the creation of fake pornography illegal, and federal efforts are ongoing. |
In addition to the evolution of these technologies and the establishment of rules, improving media literacy (the ability to critically interpret information) through school education and other means is also a crucial issue that society as a whole must address.
Summary: Discern Wisely, Embrace the Future’s Tools
So, we’ve explored the light and shadow of deepfakes. What do you think?
Deepfakes are like a “double-edged sword.” Depending on how they’re used, they can become a “tool of the future” that innovates entertainment and medicine, or a “vicious weapon” that deceives people and throws society into chaos.
The important thing is for us to be aware of this technology and cultivate a healthy critical mindset, thinking, “maybe this isn’t real either.” And to get into the habit of checking the source of suspicious information, instead of being swayed by it.
In this coming age where AI will be commonplace, I hope this article has been of some help as a first step towards wisely engaging with this new technology.
