{"id":172864,"date":"2023-09-26T16:27:51","date_gmt":"2023-09-26T21:27:51","guid":{"rendered":"https:\/\/lifeboat.com\/blog\/2023\/09\/openais-gpt-4-with-vision-still-has-flaws-paper-reveals"},"modified":"2023-09-26T16:27:51","modified_gmt":"2023-09-26T21:27:51","slug":"openais-gpt-4-with-vision-still-has-flaws-paper-reveals","status":"publish","type":"post","link":"https:\/\/lifeboat.com\/blog\/2023\/09\/openais-gpt-4-with-vision-still-has-flaws-paper-reveals","title":{"rendered":"OpenAI\u2019s GPT-4 with vision still has flaws, paper reveals"},"content":{"rendered":"

<\/a><\/p>\n

When OpenAI first unveiled GPT-4<\/a>, its flagship text-generating AI model, the company touted the model\u2019s multimodality \u2014 in other words, its ability to understand the context of images as well as text. GPT-4 could caption \u2014 and even interpret \u2014 relatively complex images, OpenAI said, for example identifying a Lightning Cable adapter from a picture of a plugged-in iPhone.<\/p>\n

But since GPT-4\u2019s announcement in late March, OpenAI has held back the model\u2019s image features, reportedly<\/a> on fears about abuse and privacy issues. Until recently, the exact nature of those fears remained a mystery. But early this week, OpenAI published a technical paper<\/a> detailing its work to mitigate the more problematic aspects of GPT-4\u2019s image-analyzing tools.<\/p>\n

To date, GPT-4 with vision, abbreviated \u201cGPT-4V\u201d by OpenAI internally, has only been used regularly by a few thousand users of Be My Eyes, an app to help low-vision and blind people navigate the environments around them. Over the past few months, however, OpenAI also began to engage with \u201cred teamers\u201d to probe the model for signs of unintended behavior, according to the paper.<\/p>\n","protected":false},"excerpt":{"rendered":"

When OpenAI first unveiled GPT-4, its flagship text-generating AI model, the company touted the model\u2019s multimodality \u2014 in other words, its ability to understand the context of images as well as text. GPT-4 could caption \u2014 and even interpret \u2014 relatively complex images, OpenAI said, for example identifying a Lightning Cable adapter from a picture [\u2026]<\/p>\n","protected":false},"author":578,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[493,1512,6],"tags":[],"class_list":["post-172864","post","type-post","status-publish","format-standard","hentry","category-climatology","category-mobile-phones","category-robotics-ai"],"_links":{"self":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/172864","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/users\/578"}],"replies":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/comments?post=172864"}],"version-history":[{"count":0,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/172864\/revisions"}],"wp:attachment":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/media?parent=172864"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/categories?post=172864"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/tags?post=172864"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}