Google's Gemini AI Now Simulates Making Calls with a Faux Buddy
Google's revamped Gemini Live AI chatbot, aptly named Gemini 2.0, is shaking up the conversation game. Despite not quite living up to the hype as an engaging conversational companion, Google wants you to imagine Gemini Live as a trusty sidekick. And with Google's latest updates, the bot is now seemingly calling you on your phone instead of you chatting with a cloud-based AI model.
In a move that's quickly catching the attention of tech enthusiasts, Google has made Gemini 2.0 Flash available to all Gemini app users, eliminating the need for a subscription. Seems there's a bit more going on behind the scenes here, as tech site 9to5Google noticed a subtle change in Gemini Live notifications. Now, when you exit the AI chatbot on Android, it appears as a call with options to "Hang Up" or put it on "Hold." If you're interacting with Gemini Live from the lock screen, you'll see a notification that "Live with Gemini" is "Listening."
Google's long-term vision for Gemini Live involves giving it vision capabilities from Google DeepMind's Project Astra assistant, allowing it to process data in real-time without pestering the user. Moreover, with the addition of multimodal learning capabilities, Gemini Live can now understand uploaded image and video content.
Google's broadening its Gemini 2.0 offerings to include smaller and larger AI models for various use-cases. The company hosted a blog post showcasing the experimental Gemini 2.0 Pro, which they claim to be their most powerful user-end model yet. Despite being geared towards coders and programmers, it will be accessible to Gemini Advanced subscribers. Meanwhile, open-source AI company, OpenAI, has been turning heads with its o3 reasoning model and its more compact o3-mini reasoning model.
Google has teamed up with Samsung for some exciting AI features on the Galaxy S25 phone. The AI can now perform simple actions like converting a text message into a calendar invite hands-free, although more complex tasks are still a work in progress. Google plans to save its best mobile AI features for the anticipated release of the Pixel 10, closer to Google I/O 2025.
But here's the lowdown on the enrichment data:
- Multimodal Output:Gemini 2.0 can generate content in images, text, and audio with native image output and multilingual native audio output with eight distinct voices and various accents.
- Agentic AI Enablement:The model provides multimodal understanding, coding, function calling, and can follow complex instructions - acting on behalf of users with their supervision.
- Native Tool Use:Gemini 2.0 can use native tools like Google Search and Google Maps as part of an LLM query or agentic AI workflow.
- Multimodal Live API:The live API allows for the integration of streaming data, such as audio and video from user screens or cameras.
- Gemini Live Enhancements:Users can now add images, files, and YouTube videos to conversations, making it easier to discuss and analyze multimedia content.
- Accessibility and Extensions:Advanced users gain access to Deep Research in the Gemini mobile app, while Gemini can also help users complete tasks across various apps.
- Gemini 2.0 Models:Gemini 2.0 Flash offers enhanced performance, multimodal input support, and improved battery life.
- Gemini 2.0 Flash-Lite:This cost-efficient model maintains the same speed and pricing as the 1.5 Flash model but offers a 1 million token context window and multimodal input support.
- Gemini 2.0 Flash Thinking Mode:An experimental mode that presents a thought process when solving complex problems.
These updates aim to make Gemini Live a more powerful and versatile AI assistant for Android devices.
In the realm of technological advancements, Google's integration of multimodal learning capabilities into Gemini Live is opening up new possibilities for AI interaction, enabling the assistant to understand and process image and video content. Furthermore, the future of artificial intelligence is becoming increasingly intertwined with technology, as exemplified by Google's development of Agentic AI Enablement, allowing Gemini Live to execute commands and tasks with user supervision.
