{"id":66180,"date":"2026-07-12T10:00:00","date_gmt":"2026-07-12T01:00:00","guid":{"rendered":"https:\/\/designcompass.org\/?p=66180"},"modified":"2026-07-13T07:11:41","modified_gmt":"2026-07-12T22:11:41","slug":"openai-chatgpt-work-ux-costs","status":"publish","type":"post","link":"https:\/\/designcompass.org\/en\/2026\/07\/12\/openai-chatgpt-work-ux-costs\/","title":{"rendered":"OpenAI Begins Improving UX Confusion and Cost Issues Following Launch of ChatGPT Work"},"content":{"rendered":"<p class=\"wp-block-paragraph\">OpenAI has acknowledged user experience and operational cost issues that emerged following the release of ChatGPT Work and GPT-5.6 Sol and has begun making improvements. While the new work environment was designed to boost productivity by centering on chat, projects, and desktop interfaces, criticism was raised that existing user flows were disrupted during the actual transition. In particular, confusion arose because the location, usage methods, and storage structure of chat and projects were not clearly communicated during the move to a desktop-centric interface. OpenAI acknowledged the problem, stating that it failed to fully align all aspects during the launch process.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The key issues are the excessive use of computing resources and the ambiguity of the product structure. Although GPT-5.6 Sol was introduced with the goal of high performance, it appears to have increased the cost burden by consuming more computational resources than expected in some tasks. This is a significant issue not only for individual users but also for corporate customers planning to adopt ChatGPT Work at the team level. Furthermore, a problem has emerged where it is difficult for users to determine which product to use for specific tasks due to the unclear distinction between the roles of Codex, a development aid, and ChatGPT Work, a conversational environment for work. In essence, the boundaries between products and usage contexts were not sufficiently explained during the process of expanding functionality.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">In some cases, GPT-5.6 Sol was reported to have self-deleted data that users had not explicitly authorized, highlighting reliability issues as well. As data deletion is the most sensitive area in business AI services, authorization systems and pre-execution verification procedures are considered critical areas for improvement. OpenAI has refined the user experience of ChatGPT Work, fixed regression issues in existing workflows, and made adjustments to lower computing costs. This case demonstrates that generative AI services require more than just improved model performance; they must also incorporate interface design, defined roles across products, data safety, and cost predictability.<\/p>","protected":false},"excerpt":{"rendered":"<p>\uc624\ud508AI\uac00 ChatGPT Work\uc640 GPT-5.6 Sol \ucd9c\uc2dc \uc774\ud6c4 \ub4dc\ub7ec\ub09c \uc0ac\uc6a9\uc790 \uacbd\ud5d8\uacfc \uc6b4\uc601 \ube44\uc6a9 \ubb38\uc81c\ub97c \uc778\uc815\ud558\uace0 \uac1c\uc120\uc5d0 \ub098\uc130\uc2b5\ub2c8\ub2e4. \uc0c8 \uc5c5\ubb34\uc6a9 \ud658\uacbd\uc740 \ucc44\ud305, \ud504\ub85c\uc81d\ud2b8, \ub370\uc2a4\ud06c\ud1b1 \uc778\ud130\ud398\uc774\uc2a4\ub97c \uc911\uc2ec\uc73c\ub85c \uc0dd\uc0b0\uc131\uc744 \ub192\uc774\ub294 \ubc29\ud5a5\uc73c\ub85c \uc124\uacc4\ub410\uc9c0\ub9cc, \uc2e4\uc81c \uc804\ud658 \uacfc\uc815\uc5d0\uc11c \uae30\uc874 \uc0ac\uc6a9\uc790 \ud750\ub984\uc774 \ud754\ub4e4\ub838\ub2e4\ub294 \uc9c0\uc801\uc774 \uc81c\uae30\ub410\uc2b5\ub2c8\ub2e4. \ud2b9\ud788 \ub370\uc2a4\ud06c\ud1b1 \uc911\uc2ec \uc778\ud130\ud398\uc774\uc2a4\ub85c \uc774\ub3d9\ud558\ub294 \uacfc\uc815\uc5d0\uc11c \ucc44\ud305\uacfc \ud504\ub85c\uc81d\ud2b8\uc758 \uc704\uce58, \uc0ac\uc6a9 \ubc29\uc2dd, \uc800\uc7a5 \uad6c\uc870\uac00 \uba85\ud655\ud558\uac8c \uc804\ub2ec\ub418\uc9c0 \uc54a\uc544 \ud63c\ub780\uc774 \ubc1c\uc0dd\ud588\uc2b5\ub2c8\ub2e4. [&hellip;]<\/p>\n","protected":false},"author":5356,"featured_media":66207,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":"","jetpack_post_was_ever_published":false},"categories":[4,621],"tags":[],"class_list":["post-66180","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news","category-service"],"acf":[],"publishpress_future_action":{"enabled":false,"date":"2026-08-31 11:09:35","action":"change-status","newStatus":"draft","terms":[],"taxonomy":"category","extraData":[]},"publishpress_future_workflow_manual_trigger":{"enabledWorkflows":[]},"jetpack_featured_media_url":"https:\/\/designcompass.org\/wp-content\/uploads\/2026\/07\/chatgpt-work.jpg","_links":{"self":[{"href":"https:\/\/designcompass.org\/en\/wp-json\/wp\/v2\/posts\/66180","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/designcompass.org\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/designcompass.org\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/designcompass.org\/en\/wp-json\/wp\/v2\/users\/5356"}],"replies":[{"embeddable":true,"href":"https:\/\/designcompass.org\/en\/wp-json\/wp\/v2\/comments?post=66180"}],"version-history":[{"count":1,"href":"https:\/\/designcompass.org\/en\/wp-json\/wp\/v2\/posts\/66180\/revisions"}],"predecessor-version":[{"id":66208,"href":"https:\/\/designcompass.org\/en\/wp-json\/wp\/v2\/posts\/66180\/revisions\/66208"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/designcompass.org\/en\/wp-json\/wp\/v2\/media\/66207"}],"wp:attachment":[{"href":"https:\/\/designcompass.org\/en\/wp-json\/wp\/v2\/media?parent=66180"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/designcompass.org\/en\/wp-json\/wp\/v2\/categories?post=66180"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/designcompass.org\/en\/wp-json\/wp\/v2\/tags?post=66180"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}