一夜生成

在失去实体经验后，用数字图像替代；死亡发生后，用照片与视频挽留；记忆模糊后，用算法修复与补全。当植物死亡，影像缺失，一夜之后，影像消散，数字图像的临时性，也对应了AI生成内容的脆弱性——它们可以随时出现，也可以随时消失，不留物理痕迹。

After embodied experience is lost, digital images take its place; after death occurs, photographs and videos are used to hold on; after memory blurs, algorithms repair and complete. When the plant dies and the image is missing, after one night, the image dissipates. The ephemerality of digital images corresponds to the fragility of AI-generated content—they can appear at any moment and disappear at any moment, leaving no physical trace.

在本作品的制作流程中，AI的角色被严格限定为“中间态生成者”。艺术家绘制起点、中点、终点三个关键帧，AI负责生成三者之间的过渡序列。这种工作方法将AI置于一个特殊的位置：它不创造核心形态，不定义边界，只负责填充人类意图之间的时间空白。

这种“填充”在结构上与当代AI图像生成技术的主流用途一致——修复旧照片、补全缺失的影像、生成从未被记录的场景。它们都是一种“生成性补偿”：用技术手段填补真实的空缺。但本作品提醒观众：这种补偿永远是一种虚构。AI生成的不是植物真实的生长过程，而是一个符合手绘边界条件的、概率性的视觉序列。它填补了图像的空缺，却无法填补真实的缺失。当晨光出现，投影消散，岩石上什么都没有留下。森林只存在了一个夜晚，而那个夜晚本身也不曾真实存在。

In the production process of this work, AI’s role is strictly limited to that of an “intermediate state generator.” The artist draws three keyframes—the starting point, the midpoint, and the ending point—and the AI is responsible for generating the transitional sequences between them. This working method places AI in a particular position: it does not create the core forms, does not define boundaries, but only fills the temporal gaps between human intentions.

This “filling” is structurally consistent with the mainstream uses of contemporary AI image generation technology—restoring old photographs, completing missing images, generating scenes that were never recorded. They are all forms of “generative compensation”: using technological means to fill real voids. But this work reminds the audience that such compensation is always a fiction. What AI generates is not the real growth process of a plant, but a probabilistic visual sequence that conforms to hand-drawn boundary conditions. It fills the gap in the image, but cannot fill the real absence. The artist conducted an experimental mourning ritual outdoors, allowing the images to linger briefly in nature; when dawn appeared, the projection dissipated, and nothing was left on the rocks. The forest existed only for one night, and that night itself never truly existed.

当AI可以生成以假乱真的图像，哀悼的对象还是那个逝去的真实吗？还是说，哀悼本身已经被转化为对图像的消费？

在当代数字文化中，死亡与失去越来越多地通过图像来中介：亲人去世后，社交账号仍被保存为数字纪念页；旧照片被AI算法修复清晰；甚至有人用生成式AI根据逝者生前照片“复活”其面容与声音。这些行为都遵循同一种情感逻辑——真实已不可追回，但图像可以制造一种“仍然存在”的幻觉。

When AI can generate images that are indistinguishable from reality, is the object of mourning still the lost real? Or has mourning itself been transformed into the consumption of images?

In contemporary digital culture, death and loss are increasingly mediated through images: after a loved one passes away, social media accounts are preserved as digital memorial pages; old photographs are restored to clarity by AI algorithms; some even use generative AI to “resurrect” the faces and voices of the deceased based on photos from their lifetime. These behaviors all follow the same emotional logic—the real can no longer be retrieved, but images can create the illusion of “still being present.”