"From AI to I" Series (2) | From Film "Da Meng" to "Taitung Slow Wave": Guo Xiancong & Wu Xiaoru Analyze the 'Taiwan Sensibility' That AI Cannot Compute

When AI can generate a photorealistic image of a Taiwanese streetscape in mere seconds, why do those intimately familiar with the land still instantly detect a sense of incongruity? The answer lies in the fact that beauty is never just about the image itself—it encompasses our understanding of the land, daily life, and cultural context.

In an era where images, designs, and texts can all be rapidly generated, what truly irreplaceable value remains for creators? The answer may not lie in more powerful tools, but in lived experiences and emotional resonance formed only through firsthand walking, seeing, and feeling.

The second installment of the 2026 Taipei New Horizon "From AI to I" lecture series invites Guo Xiancong, Visual Effects Director of the film "Da Meng," and Wu Xiaoru, curator of the "Taitung Slow Wave Project," to jointly deconstruct the elusive "Taiwan sensibility"—something even the most advanced AI computational power cannot accurately capture—from two vastly different creative frontlines.

### Guo Xiancong: Behind One Layer of Fog Lies the Reconstruction of 1950s Taipei

In the film "Da Meng," "fog" is the most central visual element throughout, appearing in over 200 of the film’s total 526 VFX shots. To achieve the perfect fog effect envisioned by director Chen Yu-hsun, Guo Xiancong and his VFX team endured immense challenges. Particularly during post-production color grading, they spent two full months solely adjusting the density and transparency of the fog on screen.

"Our goal with 'Da Meng' was to create visuals so seamless that audiences wouldn’t notice the presence of special effects at all," says Guo. His primary mission upon joining the project was to digitally reconstruct a vanished city—1950s Taipei. While fog plays a crucial role in setting the film’s atmosphere, the real challenge for the VFX team wasn’t merely "creating fog," but first understanding what Taipei actually looked like during that era.

To faithfully recreate 1950s Taipei, Guo began collaborating with the director, art department, and cinematographer from the script stage, studying historical records to grasp the postwar urban landscape, living habits, and street soundscapes. They repeatedly cross-referenced historical photos, maps, and documents—analyzing everything from building orientations and road layouts to distant mountain silhouettes and vegetation patterns. Every seemingly minor detail became critical to the success of the visual effects.

"Do you know what sound was most commonly heard in Taipei at night during that era, before cars and motorcycles were widespread? It was the tapping of wooden clogs on the pavement," Guo recalls. Postwar Taipei still bore scars—collapsed ruins awaiting reconstruction and vast undeveloped green spaces—creating an urban landscape fundamentally different from today. "When viewers see Taipei’s fog-drenched streets, what’s truly being resurrected is an entire era that has already disappeared."

Guo believes AI is merely a technical tool; ultimately, the quality of a great film depends on the director’s cinematic language and the soul of the screenplay. This highlights the most fascinating difference between AI and human creation: AI can quickly generate an image that 'looks like 1950s Taipei,' but convincing someone deeply familiar with Taiwan’s history and life that 'this truly is that time' requires extensive research, observation, and judgment.

Guo’s team also experimented with using AI to assist in aging actors’ appearances. However, the results often carried Western dataset biases, producing facial features that felt unnatural and mismatched with familiar East Asian physiognomy, requiring extensive manual correction. More importantly, if technical processing compromised the actor’s expressions or performance quality, the director would abandon even the most realistic AI-generated face swap in favor of returning to the story and character integrity. "Those of us working in VFX or technology tend to think from a technical standpoint, but a great film must always return to the director’s intended narrative, visual language, and the soul of the script."

### Wu Xiaoru: AI Can Answer Questions, But Cannot Conduct Field Research

If Guo Xiancong uses digital technology to "return" to a vanished city, then Wu Xiaoru—who earned the 2023 Japan Good Design Award for the "Taitung Slow Wave Project"—chooses the opposite path: stepping into real life as it unfolds today.

Wu has long engaged with local issues through product design, branding, and curation, frequently drawing inspiration from everyday Taiwanese life. During an early exhibition in Milan, he keenly sensed how international design discourse had long been dominated by Western knowledge systems. Yet precisely because of this, he gradually realized that the deeper creators understand their own culture and can articulate its perspective, the greater their chance to engage in genuine dialogue with the world.

"There’s one thing AI simply cannot do: field research," says Wu. He believes humanity’s most irreplaceable ability is physically entering the field. If you don’t drink tea, you won’t understand how teacups are used; without understanding coffee, you can’t design a truly functional coffee machine; and without a surfer’s bodily experience, no matter how beautiful a surfboard image appears, it may fail to meet actual needs.

This core philosophy is embodied in Wu’s participation in the New Taipei City Design Exhibition, "Life in Circle." There, he visited young residents across different districts of New Taipei, conducting field research to understand their lifestyles, then fed the collected data to AI to generate six fictional personas.

"The characters are fake, but the lives are real," Wu emphasizes. The social media records, friend interactions, and room decoration details attributed to these characters were all sampled from the authentic lives of real young people in New Taipei, then translated by AI into a shared portrait of youth life in the region.

### Taiwan Sensibility Is Life Experience That AI Cannot Feel

From the vanished 1950s Taipei in "Da Meng" to the ongoing local life captured in "Taitung Slow Wave," these two seemingly different creative approaches point to the same truth: the value of creation lies not only in what is made, but in understanding why it is made.

Wu points out that the design industry has long been expected to "solve problems," but more importantly, it must first learn to "define problems." AI can accelerate finding answers, but it cannot decide for creators what questions truly need answering. Guo also believes that as software and tools become increasingly accessible, what truly distinguishes creators is no longer technical skill, but the ability to propose unique concepts and aesthetic judgments.

Liu Li-hui, CEO of Taipei New Horizon, states: "As technology continuously reshapes the speed and boundaries of creation, people need to re-examine their relationship with land, society, and culture. From irreplicable feelings, experiences, and judgments, we must continually distill unique works and lasting cultural influence."

[Related Links] Taipei New Horizon Official Website: https://www.taipeinewhorizon.com.tw/ Taipei New Horizon Facebook: https://www.facebook.com/taipeinewhorizon/ Taipei New Horizon Instagram: https://www.instagram.com/taipeinewhorizon/