Story
September 23, 2026
ChatGPT Bets Your Voice Can Replace the Control Panel
OpenAI is betting that people will increasingly delegate digital chores by speaking, while the broader market is testing whether voice can truly become the front door to complex work. The company’s mobile push embraces that future but also acknowledges that the best assistant may need to answer in text, audio or both.
OpenAI began laying the groundwork in July, when it introduced its conversational GPT-Live model, then connected voice to desktop workflows that could handle tasks in ChatGPT’s Work and Codex areas.1
On Wednesday, that capability moved to phones. ChatGPT Voice can now take spoken instructions and act across connected apps, including checking email and summarizing files in Google Drive, whether a user is on a phone or using CarPlay.2 Plus and Pro subscribers can use the mobile Work tab to draft documents and emails, summarize Slack messages, build websites and presentations, or use a cloud browser; free and Go users get plugins and connected apps.1
The pitch is not merely hands-free convenience. In a demonstration, ChatGPT voice product lead Atty Eleti had the system examine DoorDash spending, pay $5 to reserve a pickleball court and reorganize a calendar. “In the future, we really imagine people using voice as the primary modality to interact with AI,” he said.3 More than 150 million people already use ChatGPT dictation and voice tools, according to Eleti.3
Yet OpenAI’s own design reflects a less tidy reality: people can speak faster than they type, but generally read faster than they listen. “We don’t think of this as just pure voice in and voice out,” Eleti said; the response may be text, voice or a mixture.3 Users can now switch between those modes and pick up a mobile conversation on desktop, though OpenAI still keeps chat and workspaces separate.1
The race is already crowded. Anthropic has streamlined its mobile-to-desktop handoff, while Elon Musk promoted new Grok bot tools including voice calls, voice memos, email and Slack drafts, and desktop traffic routing.
4 OpenAI’s wager is that the phone becomes the place where an assistant first hears the task — and the rest of a user’s devices become where it gets done.