ChatGPT Is Getting Chattier With ‘Advanced Voice Mode’

OpenAI startled fans earlier this year when it showed off an improved voice mode for the most advanced version of ChatGPT.

The ChatGPT advanced speech mode sounds incredibly natural, unlike the robotic voice that consumers associate with digital assistants such as Alexa or Siri. It responds in real time, adjusts to interruptions, creates laughter noises when a user makes a joke, and recognises a speaker’s emotional state based on tone of voice. (During the first demo, it sounded strikingly like Scarlett Johansson).

Chatgpt

ChatGPT Is Getting Chattier With ‘Advanced Voice Mode’

Advanced speech mode, which is compatible with the most sophisticated version of the chatbot, ChatGPT-4o, will be available to premium users beginning Tuesday. Advanced speech mode will initially be available to a small number of app “Plus” subscribers, with the goal of making it available to all Plus users by the autumn.

ChatGPT already has a less advanced voice mode. However, the introduction of a more advanced voice mode could be a watershed moment for OpenAI, transforming what was already a significant AI chatbot into something more akin to a virtual, personal assistant with whom users can engage in natural, spoken conversations in much the same way they would with a friend. The simplicity of chatting using ChatGPT’s sophisticated voice mode may inspire users to interact with the tool more frequently, posing a challenge to virtual assistant incumbents such as Apple and Amazon.

However, implementing a more advanced voice option for ChatGPT raises serious concerns: will the tool reliably grasp what users are attempting to say, even if they speak differently? Will people be more trusting of a human-sounding AI assistant, even if it makes mistakes?

OpenAI had planned to deploy advanced speech mode in June, but claimed it required “one more month to reach our bar to launch” to evaluate the tool’s safety and ensure it can be used by millions of people while still providing real-time responses.

According to a Tuesday statement, the company has recently tested the AI model’s vocal capabilities with more than 100 testers to discover any flaws, “who collectively speak 45 different languages and represent 29 different geographies.”

To avoid impersonation, the business claimed voice mode will only be able to utilise four pre-set voices produced in partnership with voice actors, and it will also block some requests that attempt to generate music or other copyrighted material. OpenAI claims the tool will have the same safeguards as ChatGPT’s text mode to prevent it from producing unlawful or “harmful” content.

chatgpt

ChatGPT Is Getting Chattier With ‘Advanced Voice Mode’

One key difference between advanced voice mode and the demo shown by OpenAI in May is that users will no longer be able to access the voice that many (including the actor herself) thought sounded like Johansson’s. While OpenAI has stated that the voice was never intended to sound like Johansson and was built with the assistance of another actor, it halted use of the voice “out of respect” after the actor objected.

The launch of ChatGPT’s advanced speech mode follows OpenAI’s announcement last week that it was developing a search engine powered by its AI technology, as the business expands its portfolio of consumer-facing AI solutions. The OpenAI search engine may someday offer a significant competitive threat to Google’s supremacy in online search.

SOURCE | CNN