cyrano@lemmy.dbzer0.com to Technology@lemmy.worldEnglish · 2 天前Microsoft Probing If DeepSeek-Linked Group Improperly Obtained OpenAI Datawww.bloomberg.comexternal-linkmessage-square77fedilinkarrow-up1409arrow-down114file-text
arrow-up1395arrow-down1external-linkMicrosoft Probing If DeepSeek-Linked Group Improperly Obtained OpenAI Datawww.bloomberg.comcyrano@lemmy.dbzer0.com to Technology@lemmy.worldEnglish · 2 天前message-square77fedilinkfile-text
minus-squareNutWrench@lemmy.worldlinkfedilinkEnglisharrow-up20arrow-down1·20 小时前When a writer copies someone else’s work without cites or compensation, it’s called “plagiarism.” But when an AI does it, it’s called “LLM training.”
minus-square𝕸𝖔𝖘𝖘@infosec.publinkfedilinkEnglisharrow-up17·20 小时前Unless that AI is not OpenAI, then it’s “plagiarism” still.
minus-squareintensely_human@lemm.eelinkfedilinkEnglisharrow-up3arrow-down1·17 小时前When a reader reads someone else’s work that’s called “reading”. But when an AI does it, it’s called “training”.
When a writer copies someone else’s work without cites or compensation, it’s called “plagiarism.” But when an AI does it, it’s called “LLM training.”
Unless that AI is not OpenAI, then it’s “plagiarism” still.
When a reader reads someone else’s work that’s called “reading”. But when an AI does it, it’s called “training”.