Google took advantage of its I/O developers conference on Tuesday, May 14, to announce new software for its search engine, its Android smartphone OS and various in-house online services. This year, the ads were entirely focused on artificial intelligence (AI), featuring Gemini, the equivalent of ChatGPT at Google.

Gemini will finally make its appearance in Google’s search engine this week in the United States, 15 long months after Microsoft itself integrated its Copilot assistant into its own search engine, Bing. This AI-powered Google is expected to be expanded to other countries by the end of the year. However, we should not expect an upheaval of the search experience, since Gemini will only intervene in response to some of the questions, when the engine judges that its intervention is appropriate.

In the coming months, the Gemini assistant will also be integrated into the heart of the Android smartphone OS. We will be able to summon this AI over numerous applications, to ask it questions, or generate images for example.

However, many areas of uncertainty remain. How we will launch Gemini, for example, remains unclear. It is unknown whether this AI will be able to speak fluently, as ChatGPT knows how to do. We also do not know at the moment how Gemini will cohabit with the old home assistant, which is traditionally summoned by saying “OK Google”.

It is difficult to predict how many smartphones will benefit from this new functionality: not all old Android terminals will receive this update and not all new Android mobiles will be immediately equipped with it.

Gemini will be integrated into Google Photos this summer, initially in the United States only. We can formulate a complex sentence to ask the AI ​​to find a specific photo, or even to extract information from it. Depending on the company, it will, for example, become possible to obtain the license plate number of our car, or the date of the first time a child managed to swim.

• Office automation

Workspace, Google’s paid office suite, is working to catch up with Microsoft’s Office, which has been integrating AI for months. Gemini will therefore soon be incorporated into the Google Meet videoconferencing software as well as Gmail messaging, but only for users registered with Google Labs, Google’s experimental test program. Google then plans to expand the user base with its AI in Workspace.

In 2025, users of its office suite working in teams will also be able to welcome a new member: an avatar of Gemini to whom they can give a name. This AI having followed all the exchanges diligently, it will be able to inform the participants about the progress of the project and answer specific questions, which perhaps constitutes the only fresh idea within the announcements regarding office automation. .

• Creation of photos, videos, music

Google presented a new version of its AI image creation tool, much improved according to the company, Imagen 3. Two creative tools were also unveiled to replace existing AI: a video generation service, now called Veo , as well as a music creation AI, now called Music AI Sandbox. These two tools are currently reserved for a handful of creative people, the only ones authorized to test them.