OpenAI announced that its AI browser experiment, ChatGPT Atlas, will cease operations on August 9, less than a year after its October 2025 launch. The company said the browsing features will be integrated into its new ChatGPT Work platform. Atlas was launched to let the browser read web pages, rewrite content, answer questions and eventually perform clicks itself. It was presented as a way to bring AI directly to the open web. Soon after launch, security researchers demonstrated that prompt injection attacks could cause Atlas to follow malicious instructions embedded in web pages. They also found that malformed URLs could expose information about previously visited sites. Neither issue caused immediate failure, but they highlighted limits of the technology. OpenAI said it will not continue a standalone browser and will instead embed its browsing functions in ChatGPT Work. The new platform bundles ChatGPT, Codex and other tools into a single desktop application. ChatGPT Work runs on GPT‑5.6 and includes plugins for Slack, Microsoft Teams, Google Drive, SharePoint, email, calendars and other business services. Users can access these tools automatically or on request. The application can generate documents, spreadsheets, presentations and websites and support long‑running projects. Codex is no longer offered as a separate product. It is now part of ChatGPT Work and adds inline diff editing, pull request reviews and support for multiple repositories. ChatGPT Work is available on desktop for all plan tiers immediately. Web and mobile versions will roll out to Plus, Pro, Business, Enterprise and Edu users over the next few days. The shift moves AI browsing from a separate browser to a component of broader productivity tools. The effect on the future of AI‑first web navigation remains to be seen.