Technology AI, Science and New Things
OpenAI’s ChatGPT has found its voice. The artificial intelligence startup will start rolling out an update to its chatbot that will allow ChatGPT to respond to questions and commands using speech. Users will be able to select from five different voices with (slightly questionable) names like Juniper, Ember and Cove. ChatGPT will then produce audio responses in that voice, creating a sense that users are having a conversation. The WSJ says at times the voice and tone of the chatbot are “almost indistinguishable” from a human – as long as the internet connection holds out. The company hopes the update will put ChatGPT in competition with features like Apple’s Siri or Amazon’s Alexa – personal assistant programmes that have a more limited language understanding but are better at performing specific tasks like playing music. Spotify, the music streaming platform, also said yesterday that it would start using OpenAI technology to translate podcasts – while staying close to the original narrator’s voice.
