Google announced two major Gemini updates on March 26. First, Gemini 3.1 Flash Live — a voice-first conversational model that can see your webcam, share your screen, and walk you through tasks in real time. It improves on Gemini 2.5 Flash Native Audio with better recognition of pitch, pacing, emotional cues, and noisy environments. It's being rolled out across Google's products including Search.
Second, under the codename 'Robin Import,' Google launched memory migration tools. 'Add Memory' captures personalization preferences from other AI services, while 'Import Chats' lets you upload exported ZIP files from ChatGPT, Claude, or Microsoft Copilot to continue conversations in Gemini.
The memory import is a strategic play to reduce switching costs. If your preferences, conversation history, and customizations travel with you, there's less friction in trying Gemini.
Google also embedded Gemini 3 into Chrome's sidebar for page Q&A and tab comparison, and launched 'Auto Browse 2' for AI Pro/Ultra subscribers — an agentic browsing feature that handles multi-step web tasks on your behalf.
Meanwhile, Lyria 3 Pro — Google's music generation model — now creates tracks up to 3 minutes with structured elements like intros, verses, choruses, and bridges. All outputs are watermarked with SynthID. Available to paid subscribers.