Last week's Google I/O conference delivered one of the company's most ambitious AI announcements in years, with new models, a redesigned interface, and deeper agentic capabilities all taking centre stage.
The headline launch was Gemini Omni, a new model capable of generating content from any input — starting with video. Users can combine images, audio, video and text to produce high-quality video output grounded in real-world knowledge.
Alongside it, Google released Gemini 3.5 Flash, positioned as a frontier model for agents and coding, with particular strength on complex, long-horizon tasks.
On the product side, Google is moving aggressively into agentic territory. Gemini Spark, a 24/7 personal AI agent integrated with Gmail, Docs, Slides and other Google tools, can continue working in the background even when a device is closed or locked. A separate Daily Brief feature pulls together urgent emails, calendar events and follow-up details into a personalized morning digest.
Search is getting a significant overhaul too — and the direction of travel will intensify debate around what The Verge's Nilay Patel termed "Google Zero": the moment Google stops functioning as a gateway to the web and fully becomes the destination. New information agents will operate continuously in the background, reasoning across the web to deliver answers directly to users. The more capable these responses become, the less reason anyone has to click through to the underlying sources — a concern publishers are already living with.
On trust and safety, Google expanded SynthID — its AI content watermarking technology — to Search and Chrome, with OpenAI, Kakao and ElevenLabs among those now adopting the standard.