Damn Multiverse Director: Stop posting comparison photos from ten years ago vs now. Your photos are being used to train AI.

Guan Jiayong Warns that AI is Devouring Visual and Creative Ecosystems, Calls on Creators to Mourn the Old Order First, Uphold Core Values, and Sow Seeds for the Next World—Don’t Let Unintentional Sharing Become Training Fuel.

As AI rapidly infiltrates the film and television industry and even rewrites the creative ecology, Daniel Kwan, director of “Everything Everywhere All at Once,” used a seemingly joking reminder during a conversation to highlight the underestimated risks of AI: “Please don’t keep posting photos from ten years ago vs. now. They’re using those photos to train machines, teaching them how humans age.”

Guan Jiayong openly states that recent social media trends of revisiting 2016 have made him reflect on a decade of dramatic change. But what concerns him more is that people upload private images as nostalgic material, ignoring the fact that these images could become fuel for model training. He warns: “Stop. Really, be careful.”

2016 to 2026: Not Just Ten Years, but the End of a World

Guan Jiayong describes “ten years” as crazy. Reflecting on his debut at the Japan Dance Film Festival with “Swiss Army Man” in 2016, winning big at the Oscars, and now returning to Japan Dance, he feels more nostalgic than celebratory.

One reason is that the Japan Dance Film Festival is about to say goodbye to its longtime home in Park City; this year is the last time it will be held there. For him, this isn’t just about relocating the festival but signifies a larger sense of ending.

This feeling of closure is compounded by signals from various industries. He mentioned on stage that Vimeo, a video creation platform, recently laid off a large number of staff and is preparing to transform into an AI company. He said, “Vimeo used to be our home.” Without Vimeo’s Staff Pick recommendations, many early works might not have been seen, and career paths might not have opened. In his view, these events all point to the same thing: the old systems of creation and distribution are stepping off the stage, while AI is rapidly taking over the new order.

Everything seems to be heading toward an end, but he chooses to see AI as a transitional phase

Guan Jiayong admits that when he looks at the world, beyond the upheavals in the film industry, even the global order is shifting; combined with the accelerated advancement of AI, everything seems to be heading toward an end. But he adds a more crucial judgment: this sense of apocalypticism is actually an entry point for understanding AI.

He says, “The way I approach AI now, or the way I can avoid being overwhelmed by it, is to understand: we are in a transition period.”

In his framework, “end” is not just about loss but about acknowledging that certain things must be mourned: old systems, old divisions of labor, old platforms, and old power structures. Only by first recognizing what is disappearing can we take the next step—protect core values and sow seeds for the next world. He states, “Only when we see clearly what is ending can we safeguard our core values and plant seeds for the next world.”

Creators in the AI Era: Mourn First, Protect, Then Sow

This “transition period” narrative has also become the underlying logic of Guan Jiayong’s recent creative and public advocacy. He emphasizes that this is not just about AI tools making filmmaking easier, but a more fundamental issue:

• What should we safeguard when the old world ends?
• What are we fighting for?
• What should we plant for the next world?

For him, AI is not just a technical issue but a competition for redistributing agency: as creators, platforms, capital, and model companies compete for discourse, without enough vigilance and collective action, rules will be written by those with the most resources. The reminder to “stop posting 2016 photos” is a modern allegory he wants to convey: in the AI era, every unconscious upload and habitual share could become training data that others use to control the future.

• This article is reprinted with permission from: 《Chain News》
• Original title: “‘Everything Everywhere All at Once’ Director Guan Jiayong: Stop Posting 2016 Comparison Photos, You’re Revealing Human Aging Models to AI”
• Original author: Neo