OpenAI has announced the alpha release of the ChatGPT's advanced voice mode, powered byinnovative GPT-4o model. This revolutionary feature, which promises to deliver hyper-realistic audio responses, will initially be accessible to a small group of ChatGPT Plus users starting today, with the aim of extending availability to all Plus subscribers by autumn 2024.
Voice Mode: ChatGPT's voice mode is official
The presentation of the voice mode with GPT-4o, which took place in May, had aroused great interest and admiration among the public, thanks to its extraordinary generation speed and the striking resemblance to a real human voice, which has evoked comparisons with that of the famous actress Scarlett Johansson.
OpenAI denied having used the actress's voice, subsequently removing the voice from the demo and announcing, in June, a postpone the release of advanced voice mode to improve security measures.
In this alpha phase, some of the features shown during the spring update, such as video and screen sharing, will not be included and will be introduced at a later time. However, a selected group of premium users will have access to the ChatGPT voice feature presented in the demo.
Read also: SearchGPT: what the OpenAI search engine is and how it works
The Advanced Voice mode differs from the audio solution currently offered by ChatGPT, which is based on three distinct models. GPT-4o, thanks to his and gentle cosmetics multimodal, is able to process these tasks without resorting to auxiliary models, generating vocal responses with one Significantly reduced latency. Additionally, OpenAI highlights GPT-4o's ability to perceive emotional nuances in the voice, such as sadness and excitement, and to recognize singing.
To ensure careful monitoring of ChatGPT's voice usage, OpenAI has opted for a gradual release. Alpha program members will receive a notification in the ChatGPT app, followed by an email containing instructions for using the feature.
The vocal capabilities of GPT-4o they have been tested in 45 languages, demonstrating OpenAI's commitment to making this technology accessible to a broad international audience. Additionally, the company has implemented new filters for prevent music creation requests or other audio material protected by copyright, thus protecting intellectual property.
ChatGPT's enhanced voice mode will initially be limited to four preset voices by ChatGPT: Juniper, Breeze, Cove and Ember, created in collaboration with professional voice actors. ChatGPT will not have the ability to imitate other people's voices and will block any output that deviates from these default voices, thus ensuring ethical and responsible use of the technology.
