OpenAI wants you to 'blurt out your thoughts' to your phone

OpenAI is launching an AI audio model that allows users to speak with their phone, according to a July announcement by the company's president, Greg Brockman. This new release integrates OpenAI's voice model with its agentic text models, which can handle tasks. OpenAI has not yet released a direct competitor to popular AI assistants like Instinct and Meta's Muse.

The company is attempting to enter the fiercely competitive market for personal agents, specifically with voice technology. OpenAI's voice model, GPT-Live, supports both listening and speaking simultaneously, which was already active in ChatGPT's conversational mode on desktop. The voice mode on mobile and the website enables users to delegate tasks, including browsing and managing apps, just by speaking.

OpenAI's product lead for ChatGPT voice, Atty Eleti, demonstrated the tool's capabilities, such as analyzing his DoorDash spending, booking a pickleball court, and reorganizing his calendar. Eleti envisions voice as the primary means of interacting with AI, aiming for people to use it as their primary modality. Personal AI assistants like Instinct and Meta's Muse are gaining traction, and OpenAI is capitalizing on the trend.

The voice tool is designed to make it easier for people to delegate tasks to ChatGPT while on the go. In the long run, OpenAI aims to leverage this technology for its hardware ambitions, including the development of a smart speaker set to release in 2027. Currently, more than 150 million users utilize ChatGPT's dictation and voice mode features.

However, the team faces challenges, as smart speakers have limited market presence, and offices still predominantly revolve around desktop computers. Despite the faster typing option, Eleti believes voice interaction offers advantages, such as convenience, multitasking, and specific benefits like language learning and public speaking practice.