(Credit: Michael Kan/PCMag)
To take on OpenAI's advanced voice mode for ChatGPT, Google is launching its own conversational feature with its Gemini chatbot.
The new feature is called Gemini Live, and it promises to let you speak to the chatbot in a free-flowing conversation rather than through a traditional input and output setup. "You can even interrupt mid-response to dive deeper on a particular point or pause a conversation and come back to it later,” Google wrote in a blog post.
Google showed off Gemini Live in a demo during its Tuesday event for Pixel 9 series. Before asking a question, the user can choose what kind of voice they want for the chatbot.
In the demo, a Google executive asks Gemini Live for tips on creating a science experiment with their child. The chatbot responds with little lag when providing its answer and follow-up responses, making the conversation seem more human-like. In addition, Gemini Live works hands-free on the user’s smartphone.
“It’s like having a sidekick in your pocket who you can chat with about new ideas or practice with for an important conversation,” Google added in the blog post.
Rick Osterloh, Senior VP of Google Devices and Services, also teased that Gemini Live will be smart enough to conduct in-depth research and produce research reports. The report will be written in a Google Doc and include sources.
Gemini Live launches today with support for 10 new voices, but the conversational mode won’t be free. Google is rolling it out in English to those with a Gemini Advanced subscription, which costs $19.99 and offers access to the company’s most powerful Gemini 1.5 Pro model. Android phones get access first; iOS support will arrive in the coming weeks.
In contrast, OpenAI only began offering its advanced voice mode last month to a small group of people with subscriptions to ChatGPT Plus, which also costs $20 per month.
The decision to paywall the voice-based chat features may disappoint users. But it suggests that Google and OpenAI are still working on ways to scale up the technology, iron out the bugs, and tackle potential legal issues before releasing the technology more widely.