{"id":4091,"date":"2023-07-13T03:44:23","date_gmt":"2023-07-13T03:44:23","guid":{"rendered":"https:\/\/www.godefy.com\/ai-loses-its-mind-after-being-trained-on-ai-generated-data"},"modified":"2023-07-13T03:44:23","modified_gmt":"2023-07-13T03:44:23","slug":"ai-loses-its-mind-after-being-trained-on-ai-generated-data","status":"publish","type":"post","link":"https:\/\/www.godefy.com\/ai-loses-its-mind-after-being-trained-on-ai-generated-data\/","title":{"rendered":"AI Loses Its Mind After Being Trained on AI-Generated Data"},"content":{"rendered":"<p>AI&#8217;s kryptonite might just be&#8230; AI. In a fascinating new paper, scientists at Rice and Stanford University found that feeding AI-generated content to AI models seems to cause\u00a0their output quality to erode. Train generative AI models \u2014 large language models and image generators both included \u2014 enough AI-spun stuff, it seems, and this ouroboros-like self-consumption will break the model&#8217;s digital brain. Or, according to these scientists, it will drive the model &#8220;MAD.&#8221; &#8220;Seismic advances in generative AI algorithms for imagery, text, and other data types has led to the temptation to use synthetic data to train next-generation models,&#8221; the researchers write. &#8220;Repeating this process creates an autophagous (&#8216;self-consuming&#8217;) loop whose properties are poorly understood.&#8221; &#8220;Our primary conclusion across all scenarios is that without enough fresh real data in each generation of an autophagous loop, future generative models are doomed to have their quality (precision) or diversity (recall) progressively decrease,&#8221; they added. &#8220;We term this condition Model Autophagy Disorder (MAD).&#8221; In other words, without &#8220;fresh real data&#8221; \u2014 translation: original human work, as opposed to stuff spit out by AI \u2014 to feed the beast, we can expect its outputs to suffer drastically. When trained repeatedly on synthetic content, say the researchers, outlying, less-represented information at the outskirts of a model&#8217;s training data will start to disappear. The model will then start pulling from increasingly converging and less-varied data, and as a result, it&#8217;ll soon start to crumble into itself. The term MAD, as coined by the researchers, represents this self-swallowing&#8230;<a href=\"https:\/\/futurism.com\/ai-trained-ai-generated-data\">AI Loses Its Mind After Being Trained on AI-Generated Data<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>AI&#8217;s kryptonite might just be&#8230; AI. In a fascinating new paper, scientists at Rice and Stanford University found that feeding AI-generated content to AI models seems to cause\u00a0their output quality&hellip; <\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[821,65,826,300,788,104,3127,447,12,898,295],"_links":{"self":[{"href":"https:\/\/www.godefy.com\/wp-json\/wp\/v2\/posts\/4091"}],"collection":[{"href":"https:\/\/www.godefy.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.godefy.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.godefy.com\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.godefy.com\/wp-json\/wp\/v2\/comments?post=4091"}],"version-history":[{"count":0,"href":"https:\/\/www.godefy.com\/wp-json\/wp\/v2\/posts\/4091\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.godefy.com\/wp-json\/wp\/v2\/media?parent=4091"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.godefy.com\/wp-json\/wp\/v2\/categories?post=4091"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.godefy.com\/wp-json\/wp\/v2\/tags?post=4091"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}