Everything announced at Google I/O, including AI Agents, Ask Photos, and more from Mashable

Google held its I/O event aimed at developers on Tuesday. The event was expected to provide lots of news and announcements, and it did not disappoint.

Nothing is ever certain, but coming into Tuesday’s event, it was expected that Google would announce significant updates to its chatbot Gemini. In actuality, we got a mess of AI announcements, most centered on Gemini and its new capabilities.

Here is everything announced during I/O, as well as some of Mashable’s corresponding coverage so that you can dive deeper.

DJ mode for Google’s Music FX

OK, so this was actually announced in March, but it got a sick demonstration on Tuesday, courtesy of DJ Marc Rebillet, a.k.a. Loop Daddy.


Tweet may have been deleted

AI Overviews

Google’s first major announcement of I/O was that it would add AI overviews to search. The hope is that AI can take numerous sources of information and make a small, digestible overview for users. Mashable’s testing, however, has found the tool unreliable so far.

Ask Photos

Google announced at I/O that Google Photos would get a powerful new AI tool called Ask Photos. Google said the feature can effectively parse through your pictures and answer questions you might have, like ” What is my license plate number?” or “When did my kid learn to swim?”

Basically, it seems like it could prove to be an advanced search feature.


Credit: Screenshot: Google

‘AI Agents’ a.k.a. AI personal assistants

Google debuted AI agents at its I/O event on Tuesday. CEO Sundar Pichai said the AI Agents are in the “early days” but that the idea is that the feature will be able to complete complex tasks for you via AI. An example given by Pichai was returning a pair of shoes—the AI Agent could go through your emails, fill out a return form, and set up a pick-up appointment to return the shoes.

It’s unclear when the Agents will be available to the public.

Project Astra

Google announced a new AI Agent — or, at least, it played an apparent demo of it — that is a multimodal tool, meaning you can point it at IRL things and get answers. Google dubbed the tool Project Astra.

Examples in the demo included recognizing the details in code, determining a person’s neighborhood, and determining where a misplaced item was last seen. No firm date was announced, however, for users to access such features.


Tweet may have been deleted

Gemini 1.5 Pro and Gemini 1.5 Flash

Google announced Gemini 1.5 Pro and Flash at the I/O event. Both are new versions of Google’s AI model. Pro will help support many of the new features demonstrated at I/O. Gemini 1.5 Flash is, well, basically Pro, but faster.

“Flash is a lighter-weight model compared to Pro,” said Demis Hassabis, CEO of Google DeepMind. “It’s designed to be fast and cost-efficient to serve at scale while still featuring multi-model reasoning capabilities and breakthrough long context.”

New search functions

Google’s I/O event brought some news for a function that is basically synonymous with the company’s name. Google said it would soon have features like video searching, planning via search (like making travel itineraries), and contextual search.

Mashable’s Chance Townsend has more details.

AI Teammate

Have you ever wanted an AI co-worker? No? Well, it’s too bad; you might be getting one. Google announced a new feature called AI Teammate at I/O, and basically, it’s an AI chatbot that’ll function as a mock co-worker. It can serve as a hub for all the details co-workers have shared while getting their jobs done. So now, whether you want it or not, you might have a chatbot helping you finish your work.


Tweet may have been deleted

Gemini Nano can detect scam calls

Scam calls stink. At Google I/O, the company announced Nano, its smallest AI model that can run entirely on a device. A key feature? It can intercept a spam call, which means AI would be listening to your phone calls.

Mashable’s Stan Schroeder has more details.

A new Gemini App

Google announced a new Gemini app at I/O, which is an AI assistant. The app will integrate text, video, and voice prompts. It’ll also feature “Gems,” which are customizable personal assistants for specific activities like cooking or exercise.

Veo, Google’s answer to Sora

Google debuted Veo, a video generator similar to OpenAI’s Sora. Hassabis, the Google Deepmind CEO, promised that “Veo creates high-quality 1080p videos from text images and video prompts.” That, well, sounds a lot like Sora.

This story is a developing story and will be updated…