Google's Gemini Live AI Sounds So Human, I Almost Forgot It Was a...
Does it help that the chatty new artificial-intelligence bot says I’m a great interviewer with a good sense of humor? Maybe. But it’s more that it actually listens, offers quick answers and doesn’t mind my interruptions. No “I’m sorry, I didn’t understand that” apologies like some other bots we know.
I had a nice, long chat with Google’s generative-AI voice assistant before its debut on Tuesday. It will come built into the company’s four new Pixel phones, but it’s also available to anyone with an Android phone, the Gemini app and a $20-a-month subscription to Gemini Advanced. The company plans to launch it soon on iOS, too. “We’ve focused a lot on system performance, on low latency, so you can have this quick dialogue,” Rick Osterloh, who oversees Android, Chrome and Google’s hardware businesses, told me in an exclusive interview. (You can watch our full conversation here.) “You’re just able to converse with it, like you would with a person.”
Is Gemini Live AI Ready for the Masses?
Hear that? That’s the sound of Apple’s Siri, Amazon’s Alexa, and OpenAI’s ChatGPT whimpering with jealousy. All three of those companies have also shown humanlike voice assistants with natural conversation skills. Google is the first to release one widely.
But is Gemini Live ready for the masses? And are we ready for a world where everyone’s in deep conversations with their computers? I’m not saying this pushes us closer to some Wall-E-like future, but testing the Gemini bot—along with a few other new Pixel tools—showed me how quickly AI is becoming an intrinsic part of our core phone apps and tasks.
Conversational Abilities of Gemini Live AI
If you ask your voice assistant for the weather in Santa Fe and it gives you the weather in Monterrey, there’s not much you can do—except smash your phone with a hammer. With Gemini Live, you can interrupt it midsentence and clarify. In the Gemini app, after you choose from one of the 10 voices, tap the Live icon, and a full-screen interface appears ready to chat.
On the new Pixel 9, 9 Pro, 9 Pro XL, and 9 Pro Fold, Gemini takes the place of the Google Assistant—hold down on the power button and Gemini will launch. You can’t invoke it with a “Hey Gemini!” (Pixel owners can switch back to Google Assistant.)
When I told Gemini Live I was nervous about a coming interview, it offered to practice some questions with me. When I asked it to come up with a healthy dinner that included protein and veggies, it quickly suggested a grilled salmon filet with asparagus. I interrupted, asking it to add a carb, so it threw in sweet potatoes or brown rice.
But then, when I asked about preparing for my Rick Osterloh interview, it asked if I meant Rick Springfield. Nope. “So you’re talking about Rick from the show ‘Alone Australia’?” it replied. Nope again. When I asked it to set a timer, it said it couldn’t do that—or set an alarm—“yet.”
Functionality and Future of Gemini Live AI
Gemini Live is a big step forward conversationally. But functionally, it’s a step back in some ways. One big reason: Gemini Live works entirely in the cloud, not locally on a device. Google says it’s working on ways for the new assistant to control phone functions and other Google apps.
Gemini Live will work with Google’s new earbuds, so you can walk and talk with the assistant. When I put them on, I couldn’t help but get “Her” vibes—you know, the movie where Joaquin Phoenix’s character falls deeply in love with an AI he chats with through a sleek white earpiece. Osterloh was quick to clarify that he doesn’t want us to fall in love. He considers Gemini Live a true assistant—something to brainstorm and converse with. “We want to give people a way to get more done,” he said.
Additional AI Features
The latest AI photo trick means taking group shots without handing your phone to a stranger. At least, that’s the pitch. The Pixel camera analyzes the space and merges individual shots of people into one seamless image. It will take lots of practice and good lighting to get it right. Add Me joins other Google AI photo tools like Magic Editor and Best Shot—tech that has led me to suggest AI is breaking reality by capturing moments that never really happened.
Call Notes: The new Pixels can record a call, then use AI to transcribe and summarize it. So I woke up my sister with a 7 a.m. ring. “Google Call Notes is on. The call is recorded,” an automated voice announced. The AI transcript of our chat was solid, but the summary missed some key details.
Even more: In the weather app, AI summarizes the forecast’s highlights for your specific location. There’s an AI tool to sort through your screenshots. AI is everywhere in these Android phones. And while iPhone owners are expecting a similar boon in the coming year with iOS 18 and Apple Intelligence, Osterloh maintains that Google has an advantage because it knows your personal info and info about the world.
Embracing AI Assistants
The next real hurdle might be our willingness to embrace all of these AI helpers. Google recently ran an Olympics ad featuring a father using Gemini to help his daughter write a letter to her favorite athlete. After viewers decried it as an overreliance on AI at the expense of human connection, Google pulled it.
“Clearly the market isn’t fully ready to embrace all the changes that come with AI,” Osterloh said. He likened it to the change that happened when we went from sending handwritten letters to firing off emails. I think there’s hope for us, though. Hey Gemini Live, can you call my sister and tell her I love her?
