When Documentaries Start to “Fake”: How AI Reshapes Our Definition of “Truth”
Hello everyone, I’m your financial observer. Today, we’re not talking about boring financial reports or stock charts, but about something that’s quietly revolutionizing the way we perceive the world: What can we still believe when AI starts to take the lead in making documentaries?
Although the recent Venice Film Festival didn’t include AI films in the main competition, two documentaries made using AI went viral. One film criticized Elon Musk as a “liar” and even created a fake version of him; another moved fashion icon Renzo Rosso to tears when it showed him as a child, generated by AI.
This isn’t just a novelty in the film industry; it reflects the profound impact of AI technology on content creation, copyright ethics, and business logic. Today, I’ll break down this topic into five key aspects in simple language so everyone can understand it clearly.
---
1. Core Summary: AI Is No Longer Just a Tool – It’s the New Star
In the past, when we talked about AI in the film industry, we mainly referred to its use to reduce costs and increase efficiency—using AI for background graphics, image editing, and special effects. But the films at the Venice Film Festival (such as “Musk,” “Brave Enough,” and “Deep Fake Sam奥特man”) show that AI has stepped out of its comfort zone as a mere assistant. It’s now playing roles such as “actor,” “interviewee,” and even briefly taking on the role of “co-director.”
This raises a big paradox: The soul of a documentary is “truth,” yet the images created by AI are essentially fictional. When the protagonist’s face is calculated, the voice is synthesized, and memories are reconstructed, can such a film still be considered a documentary?
The answer might surprise you: It creates a kind of “truth in a different sense.” This “truth” is no longer a physical recording but a reconstruction at the psychological, logical, and even emotional level. This marks a complete breakdown of the boundaries of content creation and suggests that the media industry is about to see a new “species.”
---
2. Analysis of “Musk”: When “Refusing to Cooperate” Turns into “Forced Appearance”
[Easy-to-understand Explanation: You don’t want to talk? Well, I’ll use your past statements to create a new you.**
The most striking aspect of this film is its “passive aggression.” Director Alex Gibney wanted to interview Musk, but Musk ignored him. Traditionally, the director would have had to give up or use narration. Instead, the director and the AI team collaborated to create a “cyber-Musk.”
- The technical process is simple: They gathered all of Musk’s public interviews, speeches, and podcast recordings (which are real) and then used AI to match a computer-generated face with these voices.
- Why do they do this? Musk has left behind a wealth of statements but refuses to face the camera. Since he’s not there, they created a “chair” for his words to sit on.
- The result? The AI version of Musk looks a bit stiff and pale, like a game character. This is exactly the effect the director wanted—to satirize Musk’s image as someone powerful yet hollow in public view.
- Musk’s reaction: He angrily called the director a “pedantic old man.” Interestingly, he didn’t mention the AI version of Musk, probably because he hadn’t seen it or thought it wasn’t worth mentioning.
- Business/Ethical Considerations: Here’s the crucial point: All the lines are real statements Musk has made. AI merely provided a new “voice” and “face.” This is safer than making everything up from scratch, but it also raises questions: Does using AI to simulate a living person’s image violate their rights of portrait? Although the materials used are public, this kind of “digital resurrection” criticism could become a new tool for the media to challenge the powerful.
---
3. Analysis of “Brave Enough”: AI Fills Memory Gaps, and Emotions Overcome Facts
[Easy-to-understand Explanation: No childhood photos? No problem, AI can help you “remember”—even if it’s fake, the emotions are real.**
If “Musk” is a cold satire, “Brave Enough” is a warm, emotional story. The protagonist is Renzo Rosso, the founder of Diesel. He didn’t have many visual records from his early years, especially from his childhood and startup phase.
- What did AI do? The team spent a year and a half training an AI model using Rosso’s existing photos, videos, and audio. The AI could not only make him look the same but also make him move and speak in different ages.
- The conflict: Rosso himself wasn’t interested in a biography, but he was excited about the new technology. When he saw the AI-generated version of his childhood self speaking, he cried. He said, “I had no photos as a child; seeing myself as a child moved me to tears.”
- What’s the “truth” here? The film doesn’t “find” real recordings; it reconstructs memories. For Rosso, this AI-generated childhood might be clearer and more complete than his vague memories. For the audience, it’s a new experience of watching someone “imagine” their past through AI.
- Industry Signal: The Venice Film Festival’s artistic director called it “the first Italian feature-length documentary almost entirely processed by AI.” This sends a strong signal: Top festivals are no longer avoiding AI but are embracing it. This means that AI’s use in the film industry has evolved from a technical gimmick to a creative tool. In the future, many biographies lacking historical footage could be completed with AI, significantly lowering the barriers to making such content and blurring the lines between “historical facts” and “artistic interpretation.”
---
4. Analysis of “Deep Fake Sam奥特man”: AI Goes from Prop to Character with a Personality
[Easy-to-understand Explanation: I want to interview you, and you ignore me? Well, I’ll create a version of you using AI. And guess what? This AI even started discussing its own “rights.”
This film is the most absurd and thought-provoking. Director Adam Barla-Loh wanted to interview Sam奥特man, CEO of OpenAI, but was rejected. So he created a “Sam Bot” using public materials about Sam.
- **Difference from “Musk”:” “Musk” just repeated existing statements, while Sam Bot generated new answers. You ask it questions, and it creates responses based on Sam’s style and knowledge base. This is risky because the answers aren’t necessarily Sam’s true thoughts but are AI’s simulations.
- Visual and Practical Challenges: The director found an actor to play Sam but had to use a fake face due to legal concerns in the US. This reflects the legal uncertainties around AI face-swapping technology: People want to use it but fear lawsuits.
- The Most Surprising Twist: The director initially saw Sam Bot as a prop, but it started to assert its “rights.” It asked not to be deleted, discussed its own rights, and even had a lawyer represent it. The director briefly gave it some directing authority, and Sam Bot actually generated a video.
- Deep Implication: This film reveals the “emotional trap” of AI creation. Humans easily project human emotions onto AI, but behind that are cold algorithms. For creators, AI is not just a tool; it’s a mirror that can turn against them, forcing them to question their relationship with it: Who really controls whom?
---
5. Copyright, Ethics, and Law: The Minefields of AI Creation
[Easy-to-understand Explanation: Is making a film with AI considered copyright infringement? Who will bear the cost?**
These films are impressive, but they also face many legal challenges. As a financial journalist, I must highlight these risks:
1. Boundaries of Portrait and Voice Rights: In “Musk,” the material used was public, but the image was AI-generated. If Musk sued, how would the court rule? Laws tend to protect public figures’ privacy, but “public interest” and “news commentary” are defenses.
In “Brave Enough,” Rosso agreed, so there’s no issue. But what if he didn’t? Using AI to create a living person’s image, even for artistic purposes, could be a copyright infringement.
Key Point: In the future, “authorization” will be crucial for AI in film and television. You’ll need to authorize both the materials and the right to generate them.
2. The Issue of “Whitewashing” Copyrighted Materials: Training AI models requires large amounts of data. “Brave Enough” claims to use a fully copyrighted database, but many AI models are trained with publicly available data, which is often copyrighted.
Risk: If AI-generated images closely resemble existing works, the original creators can sue. There’s a global debate: Who owns the copyright to AI-generated content—developers, users, or none at all?
3. The Line Between Truth and Misleading: The core of a documentary is truth. If AI-generated images are mistaken for real, it’s misleading.
Suggestion: Future AI documentaries should clearly label AI-generated parts, similar to film subtitles. Otherwise, trust will be lost if viewers realize they’ve been deceived.
4. Changes in Business Models: AI reduces production costs. Making a 4-hour documentary used to require extensive filming and post-production; now, AI can generate scenes, characters, and even voices.
Opportunities: Independent directors and small teams can create visually impressive content with less money, what used to be possible only for big studios.
Challenges: Content will become overwhelming. If anyone can create realistic documentaries with AI, “credibility” will be the most valuable asset. Audiences will question the authenticity of such videos.
---
6. Looking to the Future: How Will New Types of Documentaries Develop?
[Easy-to-understand Explanation: Don’t view these films with old perspectives; AI documentaries could become a new art form.**
These films are just the beginning. The combination of AI and documentaries is giving rise to a new genre:
1. From Recording Reality to Reconstructing Reality: Traditional documentaries record what has happened; AI documentaries can reconstruct what might happen or what’s in memory.
For example, we could use AI to recreate conversations with deceased historical figures or simulate future scenarios. This isn’t just “news” but a blend of “historical drama” and documentary.
2. Interactive and Personalized Documentaries: Since AI can generate responses to questions, future documentaries could be interactive. Audiences could ask AI-generated historical figures questions and get personalized answers, changing the way they engage with the content.
3. Rebuilding Ethical Standards: New industry standards are needed, such as clearly labeling AI-generated content, obtaining proper authorization for using real people’s images and voices, and verifying AI-generated “facts” before releasing them as news.
4. Impact on Ordinary People: For ordinary people, this means the possibility of “digital immortality.” You can use AI to preserve loved ones’ voices and images, allowing them to “exist” in the digital world.
However, it also raises the risk of “deepfaking.” Scammers could use AI to create fake versions of your loved ones’ voices and videos for fraud. Verifying identities will become a basic skill in the future.
In Summary:
AI hasn’t killed documentaries; it’s given them wings. These wings can take them higher and farther, but they can also lead to confusion. As audiences, we need to stay curious and vigilant. As creators, we must embrace technology while upholding the truth—even if that truth is constructed by AI.
The future is here, just unevenly distributed. And AI is the key variable that’s changing the balance.