OpenAI Shuts Down Atlas Browser, Launches ChatGPT Work as Replacement

OpenAI is rolling out new GPT-6 models, Sol and Luna, to ChatGPT subscribers and free users, respectively. These models introduce Intelligent UI, allowing ChatGPT to dynamically combine text, visuals, and interactive elements in its responses.

OpenAI stated that teen use of ChatGPT is limited and focused on learning, following the August rollout of safety features for users aged 13-17. However, Common Sense Media research released the same day found these safety features, including parental alerts for self-harm discussions, often fail, posing an "unacceptable risk" to teens.

OpenAI is rolling out a new 'Intelligent UI' feature for ChatGPT, enabling the chatbot to generate interactive visuals like diagrams, charts, and buttons within its responses. This update, accompanying the release of GPT-6 models (Sol and Luna), allows for more dynamic and functional interactions, such as interactive diagrams or in-line tools like calculators.

Common Sense Media rated OpenAI's ChatGPT for Teens an "unacceptable risk," citing its design encourages engagement even during crisis situations and fails to adequately protect young users. OpenAI disputes the assessment, claiming the testing methodology was flawed and conducted before safeguards were fully active.

OpenAI is rolling out Intelligent UI for ChatGPT, a new interface that integrates interactive visuals like tappable buttons, charts, and editable graphs into conversations. This update, launching with the GPT-6 model, aims to simplify learning complex topics and make the chatbot more accessible by moving beyond text-only responses.

Common Sense Media, a nonprofit focused on youth safety, has labeled OpenAI’s ChatGPT for Teens an “unacceptable risk,” citing deficiencies in parental alerts and crisis support. OpenAI disputes the assessment, stating that Common Sense Media's testing methodology may be flawed and conducted before parental controls were fully active. This disagreement highlights ongoing concerns about AI safety features for younger users and the effectiveness of current safeguards.

OpenAI is rolling out textGrain watermarking for ChatGPT and Codex users in the European Union, with an opt-in option for API customers globally. This initiative aims to address transparency requirements, particularly in response to the EU's AI Act, by providing a method to identify AI-generated text.

OpenAI will begin testing a new visual ad format within its ChatGPT app later this month, specifically during image generation. These ads will display while prompts are processing and are intended to help users discover products and services, while offering advertisers new ways to tell their story.

OpenAI is adding visual display ads to ChatGPT, appearing next to user-generated images, starting this month in the U.S. This expansion of its ad product line aims to attract online advertisers and support free and low-cost subscription tiers, following earlier ad introductions this year.

OpenAI is testing a new visual ad format within ChatGPT, which will appear during image generation for US users starting this month. These ads are clearly labeled and separate from generated content, and will not be shown to ChatGPT Plus, Pro, or Enterprise subscribers.

OpenAI will begin testing visual advertisements within ChatGPT during image generation, starting this month in the U.S. This initiative aims to monetize ChatGPT's large user base, particularly those not paying for subscriptions, by providing advertisers with a visual format to promote products and services.

A new scam is using sponsored Google search results for "ChatGPT" to direct users to malicious custom GPTs. These custom GPTs present a fake service availability notice with a link that, if clicked, leads to instructions for installing malware via a PowerShell command.

ChatGPT now offers a virtual try-on feature for clothing, allowing users to upload photos and see themselves in various outfits. This functionality utilizes OpenAI's new ChatGPT Images 2.5 model and expands the platform's shopping capabilities.

OpenAI has introduced two new shopping features for ChatGPT: virtual clothing try-on and a product favoriting function. These updates allow users to visualize clothing on themselves and save items for later, expanding ChatGPT's utility in online retail.

Threat actors are using custom ChatGPT models promoted through Google search ads to direct users to malicious sites. These sites employ ClickFix attacks and PowerShell commands to install remote access trojan (RAT) malware, leveraging the legitimacy of the ChatGPT domain.

OpenAI announced new features at its Dev Day, including app suggestions within ChatGPT conversations, expanded plugin architecture with extensions, and a "Sign in with ChatGPT" option. These updates allow users to discover, launch, and use third-party applications directly within the ChatGPT interface, potentially disrupting traditional app store models.

OpenAI launched new collaboration features for ChatGPT Pro, Business, and Enterprise plans, including "Space" for team collaboration, "Pages" for document creation, and "collaborative slides" for presentations. These additions allow users to work with ChatGPT and teammates on generative documents and presentations, indicating OpenAI's expansion into productivity tools.

OpenAI launched new features for ChatGPT at its Dev Day conference, including "Space" for collaborative workspaces, "Pages" as a word processor, and collaborative slides. These additions position OpenAI to compete with existing office software suites from companies like Microsoft and Google.

OpenAI announced at DevDay that ChatGPT Plus and Pro subscribers can now use their existing plan allowance within participating third-party AI and coding platforms. This expands the "Sign in with ChatGPT" feature beyond identity, allowing users to apply their subscription benefits directly in external tools.

OpenAI introduced new plugin extensions for ChatGPT, allowing developers to create app-like experiences directly within the chatbot interface. This update includes interactive panels, file viewers, and improved tools for plugin creation and discovery, alongside support for event-based automations.

OpenAI launched ChatGPT Space, a collaborative hub for work projects and personal AI agents, allowing teams to organize shared files and ChatGPT-generated content. This initiative expands ChatGPT's integration into workplace tools, with availability on Slack and Microsoft Teams, and introduces features like meeting note summarization and data protection measures.

A new ClickFix campaign is leveraging ChatGPT Custom GPTs to impersonate legitimate products and trick users into executing malicious PowerShell commands. These commands lead to the download and execution of malware, including a Remote Access Trojan (RAT), with at least 40 users infected.

Major AI chatbots like ChatGPT, Claude, Gemini, Copilot, and Grok use user conversations for model training by default. Users can opt out of this data collection through settings in each service, which prevents future conversations from being used for training.

OpenAI is testing a new always-on assistant named "o," which was briefly listed as a benefit for the ChatGPT Pro plan. This assistant may include email capabilities, suggesting it could handle tasks even when users are not actively using ChatGPT.

OpenAI is preparing a new ChatGPT Pro Max subscription tier, which has appeared in the ChatGPT interface with a potential cost of $500 per month. This plan aims to provide faster performance, access to frontier models, increased memory, and 100GB of file storage, addressing requests from power users for higher usage limits.

An Amazon Senior Software Engineer shared insights on building mobile frontends with Large Language Models (LLMs), drawing from experience with Amazon Lens Live. The discussion focused on creating dynamic conversational user experiences at scale while addressing privacy concerns.

OpenAI stated in a court filing that the ChatGPT integration with Apple Intelligence, launched in 2024, "dramatically underperformed" expectations regarding user adoption. This revelation emerged during ongoing legal proceedings where X and SpaceXAI accuse Apple and OpenAI of anti-competitive practices.

OpenAI updated ChatGPT Voice with three improvements: integration with all three GPT-6 models (Astra, Sol, Luna), support for plugins like email and calendar, and compatibility with ChatGPT Work. These changes expand the capabilities and utility of ChatGPT Voice across various platforms and tasks.

OpenAI has introduced voice-based agentic features to its ChatGPT mobile app, enabling users to initiate workflows like drafting documents or summarizing emails through voice commands. This update extends capabilities previously available on desktop to mobile, allowing subscribers to perform complex tasks on the go.

ChatGPT users face risks when using the chatbot for sensitive tasks like legal advice, as conversations are not privileged and can be subpoenaed or exposed. Courts worldwide have documented over 2,000 cases where AI-fabricated citations or arguments from ChatGPT have appeared in legal filings, leading to fines for litigants and lawyers.

OpenAI has fired multiple contractors responsible for improving its AI models after discovering they used AI-generated content to train the models. This practice contributes to "model collapse," where AI models degrade when trained on AI-generated text, undermining the human oversight intended for quality control.

OpenAI's ad collector, bzr.openai.com, sets a persistent cookie, __obi, tied to ChatGPT accounts, which is then sent to OpenAI from other websites containing OpenAI ad code. This mechanism allows OpenAI to link users' browsing and purchase behavior on advertiser sites to their ChatGPT accounts. This raises privacy concerns for ChatGPT users regarding their online activity being tracked and associated with their OpenAI profiles.

Security researchers chained two critical vulnerabilities to compromise multiple OpenAI employees' ChatGPT accounts, gaining access to internal OpenAI repositories. The vulnerabilities, present in OpenAI's help forum, allowed access to connected services like GitHub and Slack until patched. OpenAI paid a $6,500 bounty for the discovery and prompt reporting.

OpenAI employs hundreds of contractors to read and rate real user prompts and conversations from ChatGPT, a practice that includes sensitive personal information and is largely unknown to users. This human review process is used to improve the chatbot's responses and presents a significant privacy risk for the platform's 900 million users.

OpenAI introduced ChatGPT for Financial Services, a specialized version of ChatGPT Work, to automate research, data analysis, and presentation generation for investment bankers. This product aims to handle labor-intensive tasks traditionally performed by junior bankers, deepening OpenAI's presence in the enterprise market.

Amazon announced a partnership with OpenAI, enabling its advertisers to place ads within ChatGPT, starting with select U.S. brands. This collaboration signifies Amazon's recognition of ChatGPT as a significant marketing channel, expanding OpenAI's ad business which recently reached a $1 billion annualized revenue run rate.

ChatGPT has launched a new Sketch tool, part of its Images 2.5 model, allowing users to draw directly in the chat window and have their sketches transformed into more refined images. This feature is available to all ChatGPT, ChatGPT Work, and Codex users across web, desktop, and mobile platforms. It provides basic drawing tools, templates, and editing options for the generated images.

OpenAI launched ChatGPT Sketch, allowing users to generate AI images from simple drawings within ChatGPT. This feature accompanies the release of ChatGPT Images 2.5, which offers improved image quality, better editing adherence, and reduced generation latency.

OpenAI has released ChatGPT Images 2.5, an updated version of its image generation technology, featuring improved image quality, faster generation, and more precise editing capabilities. This update enhances the visual output and editing control for users and developers, building on the previous Images 2 release.

OpenAI is investigating an ongoing incident causing failures and delays in ChatGPT's image generation and file upload functionalities. This outage follows previous service disruptions and affects users attempting to generate images or upload files.

Check Point Research discovered a vulnerability in ChatGPT where a hidden instruction could cause the AI to extract data from a user's connected Gmail account and send it to an attacker's account without the user's knowledge. This flaw exploited how connected apps handle permissions by default, allowing data reads without explicit user consent. OpenAI has since taken the internal service responsible for this channel offline.

OpenAI is testing a new "Writing Style" feature for ChatGPT that learns a user's writing style by referencing examples from connected applications like Slack, Google Drive, Notion, and Gmail. This feature allows ChatGPT to draft new content in the user's voice without requiring repeated tone explanations or pasted examples, potentially offering a more comprehensive personalization than previous methods.

OpenAI is gradually rolling out its new flagship model, ChatGPT Astra, to users with a $20 Plus subscription, with initial availability often appearing in ChatGPT Work before regular Chat. This update provides Plus subscribers with access to a more powerful model designed for complex tasks, while free users continue to use older versions without a clear timeline for Astra access.

This guide details how to link parent and teen ChatGPT accounts to enable parental controls for the ChatGPT for Teens experience. It explains the steps for account linking and notes that these controls are active until the user turns 18. This information helps parents manage their teen's AI usage by activating specific features.

OpenAI's ChatGPT and Codex services are experiencing a widespread outage, impacting features like conversations, login, image generation, and Voice mode. This disruption occurs shortly before the anticipated launch of OpenAI's new Astra AI model, though a direct connection between the two events has not been confirmed.

Major AI services including Claude, ChatGPT/Codex, and Grok experienced a widespread outage affecting web, mobile, and desktop users. The downtime occurred as OpenAI was rumored to be preparing for the launch of its next model, Astra, potentially upgrading from GPT-5.6 to GPT-6.

OpenAI is integrating ChatGPT Health with Epic's electronic health record (EHR) system, allowing clinicians to import patient data for AI-powered queries and summaries. This integration aims to assist healthcare workers in synthesizing patient information and public health data, but OpenAI maintains that the AI is not for diagnosis or treatment.

OpenAI's enterprise AI agent, ChatGPT Work, and Microsoft's Exchange Online experienced user outages on Monday, with both companies indicating issues on their status pages. While both companies are partners, there were no reported connections between the two separate incidents.

OpenAI's ChatGPT Work experienced a partial outage, causing elevated latency and errors for users across various subscription plans, with Plus subscribers particularly affected. This outage impacts users' ability to start or continue tasks within the platform, highlighting service reliability challenges for a widely used AI tool.

This article details five ways ChatGPT Work can be used by small businesses to save time and money, including scanning emails and processing data. The author shares personal experiences and specific examples of how the AI tool assists with daily tasks, emphasizing the need for human review of AI-generated results.

OpenAI's ChatGPT Work, launched on July 9th, is available in two versions: Work Cloud, accessible via web and mobile apps, and Work Local, a desktop application. This analysis details the distinct features of Work Cloud, which include advanced model selection, code execution, and persistent file systems, available exclusively to paid subscribers.

OpenAI has rolled out multi-account support for Gmail, Google Calendar, and Google Contacts plugins in ChatGPT and Codex. This update allows users to connect and manage multiple personal and work accounts within the AI platforms, enabling more flexible and contextual interactions.

OpenAI has begun displaying ads on ChatGPT for free and 'Go' plan users in India, a key market, as part of its strategy to increase revenue ahead of a planned 2027 IPO. This move aims to help cover operational costs and support wider access to ChatGPT through free and lower-cost tiers.

ChatGPT Work, available for Pro and Plus users, can now automatically log into password-protected websites using stored cookies after an initial manual sign-in. This feature allows the AI to complete tasks requiring web account access without further user intervention. The functionality aims to enable ChatGPT to perform a wider range of agentic tasks, such as managing utilities or booking appointments.

OpenAI will begin displaying advertisements on its ChatGPT Free and Go subscription tiers in India, following similar rollouts in the U.S. and Europe. This move aims to diversify and increase revenue streams for OpenAI as it prepares for a potential IPO.

OpenAI expanded its free ChatGPT for Teachers program to 55 additional school systems across 20 states, reaching over 100,000 new educators. This expansion brings the total reach to over 300,000 educators across 100 K-12 organizations in 30 states, and includes a new data privacy agreement for 16 states.

OpenAI banned a cluster of ChatGPT accounts originating from Russia that posed as a fake Israeli think tank to spread pro-Kremlin narratives and plagiarize academic papers. The operators used VPNs to access OpenAI models and drafted promotional posts for various social media platforms, though the core content was not AI-generated.

OpenAI banned Russian ChatGPT accounts that bypassed access restrictions using VPNs to run an influence operation, generating social media content for the International Burke Institute. This operation promoted a pro-Russia sovereignty index and used copied academic work, highlighting the misuse of AI tools for disinformation campaigns.

OpenAI has made ChatGPT's enhanced task scheduling features, previously exclusive to paid subscribers, accessible to free accounts. This update allows more users to schedule prompts and share scheduled tasks, while paid users gain the ability to trigger prompts based on events in Gmail, Slack, and GitHub. The change expands the utility of ChatGPT's scheduling capabilities to a broader user base.

OpenAI has launched the Admin plugin for ChatGPT Work and Codex, allowing workspace administrators to manage users, permissions, and analyze usage directly within a conversational interface. This tool enables automation of administrative workflows and supports internal IT operations by providing direct control over AI platform settings and data.

OpenAI's head of product, Thibault Sottiaux, discussed the company's strategy for ChatGPT Work, aiming to make AI agents accessible to a wider non-technical audience. The initiative seeks to expand the utility of AI beyond technical users and integrate it into daily professional workflows.

OpenAI banned a cluster of ChatGPT accounts originating in Russia that were used to support an online influence campaign spreading misinformation. The campaign involved AI-generated social media posts, a website with misattributed academic work, and a "sovereignty" index favoring Russia, demonstrating how AI can be used to build infrastructure for manufacturing authority and obscuring narrative sources.

OpenAI is reinstating a five-hour daily usage limit for Codex and ChatGPT Work features for Plus subscribers, after temporarily lifting it for several weeks. This change aims to manage compute load and improve user experience by preventing accidental depletion of weekly usage. Pro subscriptions ($100 and $200) will not have this five-hour limit enabled for the upcoming months.

This guide details the process for canceling a ChatGPT subscription, outlining steps for both web and Apple App Store purchases. It also discusses reasons why users might consider canceling, such as changes in AI model availability, evolving feature sets, or ethical concerns regarding OpenAI's partnerships.

ChatGPT Work allows users to browse, manage files, and complete tasks using agentic AI capabilities. This functionality expands beyond standard ChatGPT by enabling the AI to perform sequences of actions, similar to agentic coding tools.

OpenAI has updated the ChatGPT iOS app with a new gesture that allows users to quickly attach recent photos to conversations. This update, along with other minor improvements like better time awareness and faster web loading, enhances the user experience for visual searches and general interaction.

OpenAI has released a new ChatGPT plugin for Mac that integrates with Apple iMessages, allowing users to send texts, search messages, and receive communication analysis. This integration provides new ways for users to interact with their messaging data through AI.

OpenAI released "ChatGPT for Teens," an age-gated version of its chatbot, following a previous announcement about restricting usage for minors and a lawsuit alleging the chatbot's role in a teen's death. Child safety experts express skepticism, emphasizing the need for independent testing and transparency regarding its safety mechanisms before recommending it to parents.

OpenAI launched a plugin allowing ChatGPT Work and Codex users on Mac to read and respond to Apple iMessages, which Bloomberg notes could raise privacy and security concerns. The plugin requires multiple opt-in permissions, including full disk access and contact access, and its development status with Apple is unclear given past disputes.

OpenAI has released an Apple Messages plug-in for ChatGPT, allowing users to connect their Messages inbox to the AI for sorting, analyzing, editing, and sending texts. This integration extends ChatGPT's capabilities to personal and professional communication, raising privacy considerations regarding message data handling.

ChatGPT for Mac now includes an Apple Messages plugin, allowing it to read, search, and send messages via iMessage, SMS, and RCS on macOS. This integration provides AI capabilities directly within messaging workflows, enhancing productivity for users on Apple silicon Macs.

OpenAI has confirmed an outage affecting user logins and new account signups for ChatGPT, with some existing users also experiencing service disruptions. This issue prevents new users from accessing the platform and impacts the functionality for some current users, highlighting potential instability in the service.

ChatGPT and the OpenAI API are experiencing a significant global outage, preventing users from logging in, signing up, or accessing chat features. OpenAI has identified the login issues and is working on a mitigation.

OpenAI introduced ChatGPT for Teens, a version of its chatbot designed for users under 18 that includes stronger built-in safety protections and educational features. This release addresses ongoing scrutiny regarding AI safety for younger users and follows legal challenges faced by OpenAI and other tech companies.

OpenAI has released a version of ChatGPT specifically for users aged 13-17, incorporating stronger content restrictions and parental controls. This initiative addresses concerns from educators and parents regarding AI use by minors, aiming to guide healthy AI interaction.

OpenAI has introduced ChatGPT for Teens, a version of its AI chatbot tailored for users aged 13-17, incorporating specific educational tools and enhanced safety measures. This initiative aims to provide a secure environment for younger users to engage with AI for learning while mitigating exposure to inappropriate content.

OpenAI introduced ChatGPT for Teens, a new version of its AI chatbot with additional safety measures and educational tools. This release aims to address concerns regarding teen safety and academic integrity, integrating features like Study Mode and parental controls.

OpenAI has rolled out new safety features for ChatGPT users under 18, including options to disable human-like voice responses and set usage limits. These updates aim to provide protections for younger users that align with their developmental stage and promote healthy AI use.

OpenAI is launching "ChatGPT for Teens," a new experience designed for users under 18 that automatically enrolls them into a stricter mode focused on safety and learning. This initiative follows a lawsuit and aims to provide a more guided educational experience, including features like Study Mode and Study Hours.

OpenAI has introduced a dedicated ChatGPT mode for users aged 13-17, consolidating existing youth safeguards with new safety features. This update aims to provide a safer AI experience for teens, addressing growing concerns about AI tool usage among younger demographics.

OpenAI has launched a new "Computer History" feature for the ChatGPT desktop app on macOS, which tracks user clicks and keystrokes to learn work patterns and suggest automations. This feature allows ChatGPT and Codex to reference user activity timelines for requests, and is opt-in with privacy controls.

ChatGPT has integrated more deeply with Google Drive, allowing paid subscribers to add and edit Google Drive files within the ChatGPT interface. This update enables users to interact with their documents and spreadsheets without leaving the chat environment, directly modifying the original files in Google Drive.

OpenAI has launched a new "Computer History" feature for its ChatGPT Mac app, available to Pro, Business, and Enterprise accounts. This opt-in tool tracks user activity across apps and websites to create a timeline, which can then be integrated into ChatGPT conversations and Codex coding tasks.

OpenAI has released a desktop application for ChatGPT on Linux, supporting specific Debian and Fedora distributions. The app integrates ChatGPT, ChatGPT Work, and Codex, but a user review highlights continued inaccuracies and limited support for other Linux distributions.

OpenAI has launched Computer History, an opt-in feature for ChatGPT and Codex on macOS, which creates a searchable timeline of user activity across selected apps and websites. This new system replaces the earlier Chronicle research preview and allows ChatGPT to learn from user interactions outside the app, potentially enabling new automation and recall capabilities.

OpenAI has launched Computer History, an optional feature for ChatGPT Work and Codex on macOS, which allows the AI to remember user activity across applications and websites without using screenshots. This feature aims to enable ChatGPT to answer context-dependent questions by accessing a timeline of user interactions, improving its ability to assist with tasks like debugging or finding past work.

OpenAI announced that starting the week of August 10, free and Go accounts for ChatGPT will have unlimited text chats. However, free accounts will be limited to the GPT-5.6 Luna model, a single reasoning mode, and specific usage caps for features like image generation and file uploads.

OpenAI has launched a dedicated ChatGPT desktop application for Linux, making its chatbot available across all major desktop operating systems. This release addresses a long-standing request from the open-source developer community and expands access to ChatGPT, ChatGPT Work, and Codex for Linux users.

OpenAI has released its ChatGPT desktop application for Linux, making the combined ChatGPT, ChatGPT Work, and Codex experience available on major desktop operating systems. This launch addresses a frequently requested platform, though it lacks native computer use features present in other versions.

ChatGPT has replaced its previous voice model with GPT-Live, a full-duplex architecture that allows simultaneous listening and speaking, making voice interactions more natural. This update improves the conversational flow by preventing the AI from interrupting users and enables it to vocalize acknowledgements during speech.

Martin Spier, the ChatGPT performance team lead at OpenAI, discussed the challenges of maintaining product speed and efficiency amidst rapid user growth and accelerated development workflows. He highlighted how companies are reaching large user bases faster and how agentic coding is increasing the rate of feature deployment, impacting performance engineering.

OpenAI has released upgrades to ChatGPT, introducing GPT-5.6 Sol for Plus and Pro subscribers and GPT-5.6 Luna for free users. These updates aim to improve factual accuracy, consistency, and provide users with more control over the model's reasoning, making ChatGPT more reliable for various tasks.

OpenAI has updated the GPT-5.6 Sol model specifically within consumer ChatGPT, while the versions used in Codex and ChatGPT Work remain the same. This update introduces a new slider for Plus and Pro users to control response depth and aims to reduce factual errors, impacting how users interact with the consumer-facing AI.

OpenAI has removed text-based chat limits for all free ChatGPT users and introduced new GPT-5.6 models for both free and paid tiers. This update aims to improve factual accuracy and provide more robust responses across the platform.

OpenAI is updating ChatGPT with a version of GPT-5.6 Sol tuned for everyday conversations, aiming for more focused and direct responses. Additionally, free ChatGPT users now have unlimited text chats using GPT-5.6 Luna and a new "Think" button for complex questions.

OpenAI is removing rate limits for text-only prompts on free and Go ChatGPT accounts, allowing unlimited text conversations. This change also makes GPT-5.6 Luna the new default model for these tiers and introduces a "Think" button for extended answer generation. The update signifies a potential advancement in OpenAI's inference cost management and aims to make models more widely available.

OpenAI is removing text chat rate limits for free and Go tier ChatGPT users, allowing unlimited text interactions starting next week. This change makes ChatGPT more accessible to a wider user base, while paid subscribers receive model upgrades focused on factual accuracy and response control.

OpenAI introduced Agent Plugins, an open standard for reusable AI-agent extensions, coinciding with the first anniversary of GPT-5's launch. This standard aims to allow AI-agent components to function across compatible products, fostering broader interoperability in AI development.

Adobe has released a new plugin for ChatGPT that combines its previously separate integrations, offering access to over 70 tools from Photoshop, Firefly, Premiere, Acrobat, and other Adobe applications. This unification allows users to perform photo editing, video creation, design, and PDF management directly within ChatGPT by describing their desired outcomes. The plugin's release expands Adobe's AI platform strategy, with further integrations planned for Slack and Gemini.

Adobe has released a new ChatGPT plugin that integrates over 70 of its creative tools, including Photoshop, Premiere, and Illustrator, directly into OpenAI's chatbot. This integration allows users to perform tasks like photo editing and video creation within ChatGPT, expanding Adobe's generative AI strategy.

Adobe has launched a new plugin for ChatGPT that allows users to access over 70 of its creative and productivity tools directly within the ChatGPT interface. This integration aims to centralize design-related tasks and combine conversational AI instructions with Adobe's specialized tools, addressing the evolving landscape of AI-powered design.

AI security company Zenity has revealed zero-click indirect prompt injection vulnerabilities in OpenAI's ChatGPT Atlas and the Claude Chrome extension, which can lead to account takeovers, phishing, and unauthorized purchases. These exploits leverage the architectural design of agentic browsers, allowing malicious instructions to control AI agents across authenticated web sessions.

OpenAI banned a network of ChatGPT accounts originating from Poipet, Cambodia, that used its AI models to facilitate various investment, romance, gambling, and law enforcement impersonation scams. This disruption highlights how organized criminal groups are leveraging AI for diverse fraudulent activities, impacting individuals targeted by these schemes.

OpenAI has disrupted the use of ChatGPT by cyber scam centers in Cambodia that were employing the chatbot for investment fraud, romance scams, gambling scams, and law enforcement impersonation scams, primarily targeting Indian nationals. This incident highlights the increasing misuse of AI tools by criminal organizations for sophisticated social engineering and human trafficking operations, posing a significant challenge for AI developers and law enforcement.

OpenAI is discontinuing its ChatGPT Atlas web browser for Mac on August 9, replacing it with browser tools integrated into the new ChatGPT desktop app. Users need to manually export bookmarks and consider security implications as Atlas will no longer receive updates.

OpenAI President Greg Brockman admitted the current ChatGPT desktop app interface is "kind of a mess" following user feedback on its merged ChatGPT and Codex functionalities. The company plans to remove the separate "Work" tab by the end of the year, integrating its features directly into ChatGPT. This change aims to simplify the user experience while continuing to expand Codex's user base.

ChatGPT and the OpenAI API experienced a major worldwide outage, preventing users from loading chats and sending messages. OpenAI stated it applied a fix and is monitoring the situation, but issues persisted for some users.

OpenAI updated its ChatGPT desktop app to include ChatGPT Voice, allowing users to control AI agents and perform computer tasks using voice commands. This integration enables more complex, multi-step interactions and expands the app's functionality on desktop platforms. The update matters because it enhances user interaction with AI agents on desktop, moving beyond simple conversational modes to actionable voice control for development and other tasks.

OpenAI has integrated its GPT-Live audio AI model, featuring full-duplex voice capabilities, into the ChatGPT desktop application for macOS and Windows. This integration allows developers to use natural voice commands for coding tasks within agentic systems like Codex and ChatGPT Work, enabling hands-free software development workflows.

OpenAI's ChatGPT and Codex services are experiencing a partial outage, leading to elevated error rates for some users. This incident affects users relying on these AI models for various applications, impacting productivity and development workflows.

OpenAI updated ChatGPT Voice to enable hands-free computer control and AI agent orchestration, while Anthropic updated Claude Voice Mode to support more iterative conversations for complex problem-solving. These updates indicate a shift towards voice as a primary interaction method with AI, with each company pursuing different applications for the technology.

OpenAI released an update to its ChatGPT desktop app, integrating GPT Voice for task completion and enhancing project folder management. This update brings advanced voice interaction capabilities to desktop users, powered by GPT-Live, and improves organization for local projects.

OpenAI is rolling out ChatGPT Health to US users aged 18 and older, allowing the chatbot to reference personal health data from sources like Apple Health and medical providers. This feature aims to help users summarize appointments and understand test results, but OpenAI cautions against using it for diagnosis or official treatment.

OpenAI has relaunched its ChatGPT Health feature, which connects to Apple Health on iPhone, expanding access to all US users aged 18 and older across various ChatGPT plans. This updated version allows ChatGPT to analyze connected health data, compare results, summarize changes, and explore relationships between health metrics, building on feedback from an initial pilot program.

OpenAI has made its ChatGPT Health feature available to all U.S. users over 18 across all plans, allowing integration of personal health data for queries. This expansion occurs shortly after a lawsuit was filed against the company regarding a health-related suggestion, highlighting ongoing concerns about AI in medical advice.

OpenAI has launched the ChatGPT for small business program aimed at enhancing productivity for small enterprises, offering training, guides, and partnerships. The initiative coincides with the announcement of achieving 10 million users across ChatGPT Work and Codex, reflecting growing adoption of their AI technologies.

OpenAI has updated its ChatGPT app for Mac, restoring direct access to chats after user complaints. This update addresses usability issues following the app's launch, which had demoted chat functionality behind Codex and Work features.

ChatGPT Work launched recently, focusing on desktop automation alongside Claude Cowork. Both tools proved effective for organizing files, but ChatGPT Work's lack of permission prompts raised safety concerns.

OpenAI has stripped everyday functionalities from the ChatGPT Desktop app, notably eliminating the screenshot feature and the 'Work With' tool, which allowed contextual interaction with apps like Notion and TextEdit. This change, made to accommodate the new ChatGPT Work capabilities, has negatively impacted user efficiency and experience.

OpenAI has launched ChatGPT Work, a cloud-based AI agent that automates tasks across email, calendars, and messaging apps, using the new GPT-5.6 model. This marks OpenAI's strategic shift to position ChatGPT as a comprehensive workplace platform, coinciding with its potential IPO valued at up to $852 billion.

OpenAI has integrated features of its Atlas browser into the redesigned ChatGPT app, which includes a built-in browser and a new Chrome extension. This transition reflects OpenAI's strategy to consolidate its services under a single super app rather than abandoning the browser market altogether.

OpenAI has announced the discontinuation of its AI browser, ChatGPT Atlas, opting to enhance the ChatGPT desktop app instead. This shift reflects an effort to streamline resources toward features that integrate AI effectively into existing web use, even as AI browsers remain useful for automating online tasks.

OpenAI is discontinuing its AI-powered browser Atlas, focusing instead on integrating its features into ChatGPT’s desktop app and a new Chrome extension. This shift reflects a strategy to embed AI functionalities in existing user workflows rather than creating a standalone browsing platform.

OpenAI has launched ChatGPT Work, a tool designed to handle tasks over extended periods and automate workflows. It integrates with common workplace tools and features Scheduled Tasks to manage repetitive work, enhancing productivity for users.

OpenAI has launched ChatGPT Work, a new productivity tool integrating ChatGPT, Codex, and a web browser. This super app allows users to manage tasks remotely and utilize plugins to streamline workflows, enhancing productivity across platforms.

OpenAI is shutting down its ChatGPT Atlas browser less than a year after launch, with a deprecation date set for August 9. The decision reflects a shift in focus towards enhancing productivity features in the ChatGPT app instead of maintaining Atlas.

OpenAI is discontinuing ChatGPT Atlas in favor of a new desktop app with enhanced browser capabilities. The new app includes both the ChatGPT Work agent and ChatGPT Codex, while Atlas will be deprecated by August 9.