Form: Variance / Convergence

Form: Variance / Convergence is an ongoing photographic project that combines cosplay, typological photography, and generative artificial intelligence to investigate how contemporary standards of beauty converge under the combined influence of digital culture and algorithmic systems.

The project originated from a simple observation: when different individuals portray the same fictional character, they appear to retain their own age, gender, physical characteristics, and personal histories, yet simultaneously move toward a remarkably similar appearance. Hairstyles, makeup, facial expressions, body language, and even methods of image retouching gradually align into a recognizable visual pattern.

This phenomenon was not created by artificial intelligence. Long before generative AI emerged, popular culture, character design, social media, and commercial imagery had already taught people what an “ideal appearance” should look like. However, with the arrival of generative AI, this process of convergence has become accelerated, amplified, and increasingly systematized. Differences that once existed between individuals are compressed into a limited set of preferred visual traits: large eyes, high noses, small mouths, smooth skin, pointed chins, and youthful faces aligned with dominant contemporary aesthetics.

Form: Variance / Convergence is not primarily concerned with whether AI will replace humanity, nor does it treat AI as merely a technological tool. Instead, the project examines a deeper cultural condition: what happens when humans and algorithms begin pursuing the same ideal image.

The green-screen studio that appears throughout the project is not simply a technical backdrop. Within Form: Variance / Convergence, the green screen functions as a visual metaphor for latent space—an intermediate territory between reality and fiction, between human subjects and machine interpretation. Within this space, people exist simultaneously as photographic subjects and as data awaiting transformation.

The project currently consists of several interconnected bodies of work.

Human Matrix presents typological grids of different individuals portraying the same character, revealing how people voluntarily converge toward established aesthetic norms.

AI Matrix places human portraits alongside AI-generated interpretations, exposing how algorithmic systems further compress difference and translate diverse human bodies into increasingly standardized faces.

Partial Erosion focuses on moments where the boundary between human and generated image begins to collapse, visualizing friction, mistranslation, and disappearance within the process of algorithmic interpretation.

Perfect AI represents the system’s imagined ideal form: a figure that appears convincingly human yet no longer belongs to any specific individual.

Moving between typological photography, identity performance, and generative imagery, Form: Variance / Convergence traces an ongoing cultural transformation in which people no longer use images merely to represent themselves, but increasingly reshape themselves according to the logic of image-making systems.

Ultimately, Form: Variance / Convergence is less concerned with what artificial intelligence generates than with what it chooses to preserve—and what it chooses to erase.

《Form：差異 / 趨同》
《Form：差異 / 趨同》是一項持續進行中的攝影計畫，透過角色扮演（Cosplay）、類型學攝影與生成式人工智慧，探討當代審美如何在數位文化與演算法系統的共同作用下逐漸趨同。

計畫最初源自一個簡單的觀察：當不同的人扮演同一個角色時，他們看似保留了各自的年齡、性別、身體條件與生活背景，卻同時朝向某種相似的外貌與姿態收斂。髮型、妝容、表情、肢體語言乃至於修圖方式，都逐漸形成一種可辨識的共同模式。

這種現象並非由人工智慧所創造。在生成式AI出現之前，人們便已透過流行文化、角色形象、社群媒體與商業影像，不斷學習何謂「理想的外貌」。然而，當生成式AI開始參與影像生產時，這種原本緩慢且分散的收斂過程被進一步加速、放大與制度化。原本存在於個體之間的差異，逐漸被壓縮為少數被系統偏好的視覺特徵：大眼睛、高鼻樑、小嘴、光滑皮膚、尖下巴，以及符合當代主流審美的年輕臉孔。

《Form：差異 / 趨同》並不試圖證明AI是否會取代人類，也不將人工智慧視為單純的技術工具。相反地，本計畫關注的是一種更深層的文化問題：當人類與演算法共同追求相同的理想形象時，個體差異將如何被重新定義，甚至逐漸消失。

作品中的綠幕攝影棚並非僅是一種技術性的拍攝環境。在《Form：差異 / 趨同》中，綠幕被視為一種「潛在空間（Latent Space）」的視覺隱喻——一個介於真實與虛構、人類與機器之間的中介場域。在這個空間中，人物既是被拍攝的主體，也是等待被重新解讀的資料。

本計畫目前由數個相互關聯的系列組成。

真人矩陣（Human Matrix） 以類型學方式排列不同個體對同一角色的詮釋，揭示人類如何主動向既定審美收斂。

AI矩陣（AI Matrix） 將真人肖像與人工智慧生成結果並置，呈現演算法如何進一步壓縮差異，並將多樣的人類身體轉譯為趨於一致的理想面孔。

局部侵蝕（Partial Erosion） 聚焦於真人與生成影像之間的邊界崩解，呈現資料轉譯過程中的衝突、誤讀與消失。

完美AI（Perfect AI） 則描繪系統最終所想像的理想形態——一個看似真實、卻不再屬於任何具體個體的人物。

從類型學攝影、身份表演到生成式影像，《Form：差異 / 趨同》試圖描繪一種正在發生的文化轉變：人類不再只是使用影像來呈現自己，而是逐漸依照影像系統的邏輯重新塑造自己。

最終，《Form：差異 / 趨同》所關注的並非人工智慧生成了什麼，而是人工智慧選擇保留了什麼，又刪除了什麼。