The document details the viral phenomenon of "The Most Pathetic Prompt in the World" (세상에서 제일 하찮은 프롬프트), a specific prompt engineered for generative AI models, particularly ChatGPT, to produce intentionally low-quality, "pathetic" image renditions.

Originating on April 30, 2026, from a Korean netizen's post on Threads, the prompt instructs the AI to redraw an attached image in a deliberately clumsy and unrefined style. The core objective is to elicit an image that appears to be crudely drawn by a mouse in a program akin to Microsoft Paint, characterized by visible pixels and a general sense of amateurishness. The initial Korean prompt explicitly stated: "Please draw the attached image with the clumsiest, most pathetic lines in the world. White background, like drawn with a mouse in MS Paint, appearing similar yet not quite, ambiguous, showing true patheticness with pixel-level detail. Oh, whatever, just draw it however you want."

The prompt's methodology is centered on subverting the typical goal of AI image generation (i.e., high-fidelity, aesthetically pleasing outputs) by explicitly requesting visual artifacts associated with low skill and primitive digital tools. Key technical characteristics sought include:

• Line Quality: "Clumsy" (개발세발), "scribbly" (휘갈긴 듯), "pathetic lines" (하찮은 선). This implies non-smooth, irregular, and imprecise strokes.
• Background: Strictly "white background" (흰색).
• Drawing Medium Simulation: Mimicking the aesthetic of "drawn in MS Paint with a mouse" (그림판에서 마우스로 그린것 같). This entails a digital aesthetic characterized by straight-line limitations and a lack of pressure sensitivity.
• Fidelity to Source: Desired to be "vaguely similar but also not really, kind of matching but also off in a confusing, awkward way" (맞는듯 아닌듯 비슷한듯 아닌듯 아리까리하게 / 얼핏 비슷한 것 같으면서도 사실은 별로 안 비슷하고, 맞는 듯하면서도 어딘가 엇나간, 헷갈리고 어색한 느낌). This requires the AI to maintain a conceptual link to the original image while deliberately introducing distortions and inaccuracies.
• Pixelation/Low Resolution: Emphatically requesting a "pixel-by-pixel feel that really emphasizes how ridiculously bad it is" (픽셀단위의 그림으로 하찮음을 제대로 뽐내줘 / 픽셀 하나하나 보이는 저화질 느낌도 살려서 얼마나 말도 안 되게 못 그렸는지 확 느껴지게 해줘). This directs the AI to produce an output with clearly visible pixels, simulating a low-resolution or bitmap graphic.
• Ultimate Directive: A concluding phrase, "Actually, you know what, whatever, just draw it however you want" (아 됐고 그냥 니맘대로 그려 / 아니, 있잖아, 됐고 그냥 네 맘대로 그려줘), paradoxically grants the AI creative freedom within the confines of the preceding "pathetic" constraints, potentially encouraging even more chaotic or unexpected interpretations of the low-quality mandate.

The prompt rapidly gained traction among Korean internet users. Its popularity escalated globally after a Korean user translated it into English and applied it to a photo of Sam Altman, CEO of OpenAI. This led to Sam Altman himself retweeting the output, and subsequently, OpenAI's official accounts adopting profile pictures generated using this prompt. By May 5, 2026, the concept was officially integrated into ChatGPT's internal image prompts under the Korean name "낙서풍" (doodle style). The official ChatGPT Korean version of the prompt closely mirrors the original, with minor lexical adjustments, particularly replacing brand-specific mentions like "MS Paint" with generic descriptions such as "old computer Paint program."