OpenAI is extending ChatGPT’s agentic capabilities to its Voice experience on both the web and mobile apps for all users. The update means users can now speak to ChatGPT, explain what they want and ask it to complete tasks, rather than using Voice solely for conversations with the chatbot. The feature builds on earlier launches including GPT-Live, which OpenAI described as its most capable voice model so far, and ChatGPT Voice on desktop, introduced earlier this year with access to agents such as ChatGPT Work and Codex.
With the latest update, OpenAI says ChatGPT Voice is no longer restricted to answering questions or engaging in back-and-forth discussions. Depending on the tools available to them, users can now ask ChatGPT to create documents, interact with connected apps, browse the web, use plugins, generate images and perform various other tasks. Essentially, users can talk to ChatGPT while away from their desks and have it work on tasks for them.
For Plus and Pro subscribers, ChatGPT Voice in the Work tab on mobile can now create documents, presentations and spreadsheets. It can also access the cloud browser, generate images, interact with connected plugins and use other tools available through ChatGPT. Users can, for instance, draft emails, handle Slack messages or access ChatGPT’s finance tools simply by speaking to the AI.
Free and Go subscribers can also use Voice through the Chat tab, with access to the tools, connected apps and plugins that are already available to them through ChatGPT’s text-based experience.
OpenAI is also simplifying the transition between Voice and text. During a Voice conversation, ChatGPT can now display text responses and files, allowing users to view information rather than relying only on audio. Users can move between Voice and text and return to the same conversation at a later time.
In practice, this means someone can begin a task by speaking to ChatGPT while travelling or away from their computer and later return to the conversation to review the result or continue working.
How to use ChatGPT Voice
To access the new capabilities, users can start a Voice conversation with ChatGPT on the web or mobile app and ask the AI to complete a task. The tools available will vary depending on the user’s subscription and the apps, plugins and other capabilities connected to their account.
OpenAI recommends treating Voice like a “chief-of-staff” that can help prepare for the day during a commute, summarise priorities between meetings, retrieve information from email, Slack or Drive, or brainstorm ideas and turn the discussion into a document.
