Meta AI voice chat feature was finally introduced on Wednesday, after rumours about it appeared earlier this week. The social media giant announced multiple new artificial intelligence (AI) features for its native chatbot at the Meta Connect 2024, among which the voice chat feature was a major highlight. It will allow users to have a two-way conversation with the AI without needing to type or read text. However, it is not an emotive voice and instead the experience is similar to text-to-speech tools.
In a newsroom post, the social media giant detailed the feature. It was also showcased during the event. Meta highlighted that the voice chat feature will be rolled out to Meta AI on Messenger, Facebook, Instagram DM, and WhatsApp.
Put simply, this is a speech mode for the AI, similar to what companies like OpenAI and Google are already offering with their chatbots. However, unlike the Advanced Voice Mode on ChatGPT, which has a human-like expressive voice, can react to user’s words, and can interact in real-time, Meta AI voice chat is more of a read-aloud experience of generated text.
Based on the demo shared by the company, the voice is not entirely robot-like, but there is little to no voice modulation. One upside, however, is that it is available to all users for free. Meta is also planning to introduce several new voice options for users to choose from. This will also include five celebrity voices including Awkwafina, Dame Judi Dench, John Cena, Keegan Michael Key, and Kristen Bell.
A report earlier this week revealed that Meta was planning to introduce celebrity voices for this feature. The company is said to have secured deals with them to use their voice for Meta AI. As such, more celebrities can also be added in the future.
To use the voice chat mode, users need to go to the Meta AI interface on any of the Meta-owned platforms that support it. There, users can find a waveform icon placed next to the text field. Tapping on it will open a window with the Meta AI icon placed at the centre and icons for speaker, mute, and messages placed at the bottom. Once here, users can begin speaking their queries and prompts and the AI will respond verbally in the chosen voice.