Updated
Updated · TechCrunch · Jul 24
OpenAI Adds ChatGPT Voice to Desktop App, Enabling AI Agent Control on Computers
Updated
Updated · TechCrunch · Jul 24

OpenAI Adds ChatGPT Voice to Desktop App, Enabling AI Agent Control on Computers

3 articles · Updated · TechCrunch · Jul 24

Summary

  • Thursday’s update lets users speak to the ChatGPT desktop app to control AI agents and carry out tasks on their computer, extending voice beyond conversation into action.
  • The feature runs on OpenAI’s new ChatGPT-Live voice models and works with both ChatGPT Work and Codex, including computer-use skills that can navigate websites and apps.
  • On macOS, Appshots can grant the app access to on-screen content, including alt-text, while users can also answer follow-up prompts when multi-step commands need input.
  • A demo showed a developer issuing one spoken request to open a new thread, create a pull request and trace a bug’s root cause; Codex voice can also be accessed remotely from the iOS app.
  • The move makes desktop voice more capable than the smartphone version launched earlier, and comes as Anthropic also expands Claude voice to complete tasks across apps like Gmail, Slack and Notion.

Insights

Will OpenAI’s screen-reading voice agent revolutionize coding, or is it a massive privacy breach waiting to happen?
As AI shifts to executing tasks, could your next coworker be a fully duplex voice model fixing bugs in the background?
With voice agents now controlling desktops and reading hidden screen data, are we witnessing the end of traditional graphical user interfaces?