自2022年以來，生成式人工智慧圖像工具逐步向公眾開放，因其具備接近「真實照片」的視覺效果，隨即引發廣泛討論。這些影像是透過演算法處理和視覺數據資料庫訓練產生，並非捕捉實際場景，與現實世界並無直接的物理聯繫，這也使人重新思考在人工智慧技術背景下攝影的價值與未來發展。由於攝影的定義與認知持續變動，人工智慧生成圖像技術引發的討論展示了攝影發展如何促使我們反思媒介技術的文化意涵。本文認為，透過探討這些「攝影事件」，可以進一步理解當下攝影在技術與媒介層面的轉變。本文以2023年德國藝術家艾爾達格森（Boris Eldagsen）以〈偽記憶：電工〉參加索尼世界攝影大獎所引發的爭議為例，並以傅拉瑟（Vilém Flusser）的理論為框架，從三個面向進行討論：數據庫作為記憶外部化的角色、提示書寫在影像生成中的作用，以及裝置（機具／機制）概念如何擴展我們對攝影的理解。

Since 2022, generative AI image tools have gradually become available to the public, initiating widespread discussion due to their ability to produce visuals that closely resemble “real photographs.” These images, generated through algorithmic processing and training on visual datasets, have no direct connection to the physical world. This has led to a reconsideration of the value and future of photography in the context of AI technology. Discussions surrounding AI-generated imagery reveal how the development of photography compels us to reconsider the cultural implications of media technology. This paper argues that by examining the “photographic incidents,” we can better understand the current transformations in photography at both technical and media levels. Taking the controversy sparked by German artist Boris Eldagsen’s entry Pseudomnesia: The Electrician in the 2023 Sony World Photography Awards as a case study, this paper applies Vilém Flusser’s theoretical framework to explore three aspects: the role of databases as externalized memory, the function of prompts in image generation, and how the concept of apparatus/dispositif expands our understanding of photography.