You can now ask ChatGPT to check your calendar and draft emails by voice
iPhone users will be able to use the capability, too.
- OpenAI added ChatGPT Voice to its desktop app for paid users, enabling hands-free task coordination.
- Users can ask ChatGPT to start tasks, check progress, change direction, and use screen context during conversations.
- OpenAI warns appshots may capture visible images and accessible text, including content outside the visible scroll area.
- The rollout comes as AI assistant competition grows, including Anthropic expanding Claude voice mode with workplace app integrations.
Key Takeaways by nexos.ai, reviewed by Cybernews staff.
OpenAI has updated its ChatGPT desktop app with the ChatGPT Voice feature, which lets you coordinate tasks, check progress, or change direction without having to use your keyboard.
The company said on Thursday that ChatGPT Voice is available in the ChatGPT desktop app with ChatGPT Plus, Pro, Business, Edu, and Enterprise plans. Enterprise and Edu customers will get two weeks of early access to the feature before it becomes available by default.
iPhone users will also be able to use the new capability by pairing their phone with a desktop host.
ChatGPT Voice can be activated by tapping “Start new voice chat” in the ChatGPT desktop app. You will need to allow microphone access, choose a voice, and review screen context on macOS during your first chat.
The company says that the feature acts much like a human conversation partner. You can interrupt it or ask a question, or request it to switch direction while it's already working on a task.
Rather than simply talking to you, ChatGPT Voice can coordinate multiple assignments – from starting tasks to checking on existing ones and sending follow-up instructions.
ChatGPT can also capture the active application window ("appshot") and analyze what's displayed on the screen to use it as context during conversations. You can activate the feature by accessing screen context in settings and saying “Take a look at this.”
From a privacy standpoint, it’s worth noting that the appshot might include not only the window’s image but also accessible text, including content outside the visible scroll area. OpenAI recommends users avoid sharing windows that contain sensitive information, even if not immediately visible in the scroll area.
Stay updated with our latest stories and follow us on social media
Be the first to discover new stories, ideas, and updates from our team.
The feature follows the same permissions as Chat, Work, and Codex, and organizations can disable screen context if necessary.
OpenAI says that only one voice chat can be active at a time. These chats have their own allowance limits, measured in rolling five-hour windows. Tasks started through Voice count toward users' Codex usage budget.
The feature is rolling out globally amid escalating competition in the AI assistant market. Earlier this month, Anthropic expanded Claude's voice mode, now integrated with workplace apps such as Gmail, Slack, and Notion.
