news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

whip bulls reported that on october 2, according to foreign reports, openai is opening its voice ai engine to other developers, which provides support for chatgpt’s advanced voice mode.

developers will have real-time access to the technology, where the ai ​​can understand voice commands and conduct voice conversations in live phone-like scenarios.

the process previously required developers to go through at least three steps: first transcribing the audio, then running the generated text model to derive an answer to the query, and finally using a separate text-to-speech model.

the move paves the way for a wave of artificial intelligence applications that offer conversational voice interfaces.

the new speech-to-speech feature is one of several announcements openai made at its devday event in san francisco on tuesday.

early testers of the feature include nutrition and fitness app healthify and language learning app speak.

other new features available to developers include the ability to fine-tune models based on images.

in a demo for reporters, openai executives showed off an example of the new audio feature combined with twilio's api, which allows an ai assistant to call a fictional candy store and order 400 chocolate-covered strawberries.

among the customization demos of the tool was one example of talking to an ai system to help find local products, such as strawberries. the ai ​​then calls the merchant to order strawberries and takes instructions from the user on how much to order and how much they expect to spend.

openai says anyone using such technology is not allowed to hide that it is artificial intelligence and not a human, and only offers six presets to developers rather than creating new sounds.

developers can only use sounds provided by openai - the same options as in chatgpt.

while the sounds aren't watermarked in any way and developers don't have to have themselves recognized by ai systems, openai says using its systems to spam or mislead people violates the company's terms of service.

the announcements come amid a flurry of news surrounding the chatgpt maker, including its ongoing massive fundraising campaign and the departure last week of chief technology officer mira murati and two other executives.