“Gemini”, an artificial intelligence from Google

Google launched its new artificial intelligence (AI) model, Gemini 1.0, on Wednesday, December 6. For the Californian giant, it’s a question of competition from OpenAI (creator of ChatGPT), funded primarily by Microsoft.

Google DeepMind Vice President Eli Collins highlighted Gemini’s strengths during a press presentation: “ It is our most substantial, most talented, and also most general model of artificial intelligence. »

This AI, which is meant to compete with Chat-GPT4, is multi-modal, meaning it will consider text, audio, images, video and be capable of logic.

Since the launch of ChatGPT a year ago, there has been a furious race between Silicon Valley giants for so-called generative AI.

Sundar Pichai, the head of Google, described the level of analysis of Gemini in a video. If the user shows Gemini objects, drawings and videos, the AI ​​verbally comments on what it “sees”, identifies objects, plays music and answers questions by justifying its “reason”.

For example, when faced with an image of a rubber duck that must choose between two paths—the one on the left that leads to another duck drawn on paper, and the one on the right that leads to a menacing-looking bear—Gemini will suggest the left path because “ It’s better to have friends than enemies “.

Google launched its own chatbot (a computer program that simulates human conversation), Bard, but remained little known compared to its rival ChatGPT. Its new Gemini model aims to improve on that. ” This is the next step towards our vision: to bring you the best AI collaborator in the world ,” insisted Sissie Hsiao, Google’s vice president in charge of Bard.

A bard must gain abilities, but always with written requests and only in English. You’ll have to wait until 2024 for other features and formats, such as advanced math help.

Read alsoArtificial intelligence: Google opens its conversational robot Bard to 180 countries

Leave a Comment