OpenAI Bets on Audio and Prepares a Separate Device with Voice Control
OpenAI is expanding its work on audio models and forming a unified team of engineers, product managers, and researchers. According to The Information, the company is preparing to launch a personal device with a voice interface.
What is Known
In the US, voice assistants have already become common thanks to smart speakers. Meta has added five microphones to Ray-Ban glasses, allowing better hearing in noisy environments. Google has been testing "Audio Overviews" since June — a feature that voices short summaries of search results. Tesla is integrating the Grok chatbot from xAI into its cars, which manages navigation, climate, and other functions via voice.
Simultaneously, startups are working on audio devices. Developers of Humane AI Pin spent hundreds of millions of dollars before their screenless gadget became a risky launch example. The Friend AI pendant, which constantly records audio and offers the role of a virtual friend, has raised concerns over privacy risks. Several companies, including Sandbar and a startup by Eric Migicovsky (founder of Pebble), are preparing to launch smart rings with voice control by 2026.
According to sources, the new OpenAI audio model will be able to speak more naturally, understand interruptions in conversation, and even respond simultaneously with the user. The company is also considering creating a family of devices — possibly including glasses or smart speakers without screens. Jony Ive, former chief designer at Apple, is also involved in OpenAI's audio projects. One of his priorities is to reduce screen dependency through new device formats.
Source: The Information
