The Californian firm presented its new artificial intelligence model on Wednesday, December 6, which would exceed the capabilities of human experts in several areas.
The race for AI is accelerating. Just a year ago, Sam Altman, head of Open AI, launched ChatGPT, a conversational intelligence capable of finding answers to many questions. Technology had shifted our world into a concrete projection of what a future surrounded by “super-intelligences” could be, bringing at the same time its share of questions about the limits to be set and their influence, particularly on work. This Wednesday, Google ignored these questions and decided to further accelerate the development of these tools by launching Gemini, a new model supposed to make ChatGPT outdated. “It is our most consistent, most talented and also most general AI model,” assured Eli Collins, a vice-president of Google DeepMind, the AI research laboratory of the Californian group, during a presentation to the press. Sundar Pichai, the firm’s CEO, clarified Google’s interests in an announcement video: “We have always seen our mission as having no time limit. We must be able to organize the information of this world and make it useful and universal for everyone.
Let’s go hands-on with #GeminiAI.
Our newest AI model can reason across different types of inputs and outputs — like images and text. See Gemini’s multimodal reasoning capabilities in action ↓ pic.twitter.com/tikHjGJ5Xj
– Google Google) December 6, 2023
“It’s a duck!”
Concretely, Google is seeking to develop omniscient AI. A technology capable of seeing everything, hearing everything, understanding everything – not just text as is the case for ChatGPT, but also images, code and sound – and then analyzing it for its purposes. users. In demonstration videos posted on YouTube this Thursday, we see a man drawing a rounded line on a post-it on camera. Gemini responds to him with a fluid, barely robotic voice: “It’s a trait. The contours are smooth and fluid without angles.
The man then draws eyes, legs and water. “It looks like a bird to me. The bird swims in the water. It has a long neck and a beak. It’s a duck! A duck is a type of waterfowl in the Anatidae family. Ducks are closely related to geese and swans.” He is then able to translate it into all languages and place it on a map. A test which may seem trivial, even almost childish, but which in reality enlightens us on the potential of this intelligence capable of seeing what we see, of commenting on it and justifying its reasoning. “In the tests we have carried out on specific subjects – particularly in industry – Gemini is as strong as the best experts on the subject,” assures Eli Collins.
“Our vision”
The deployment of Gemini already appears to be revenge for the Mountain View multinational. A year ago, she was taken by surprise by ChatGPT, whose success had been as phenomenal as it was unexpected. The secret development of this technology, more powerful in many ways before its launch in an emotional video spot just days after its rival’s birthday, looks like an answer. Google now has many assets to become the leader in the AI market. The company will provide access on December 13 to a first version of its technology to its customers in the cloud (remote computing), including developers who use its Vertex AI platform to create their own AI applications.
The Google Pixel 8 Pro, released in October, will be the company’s first smartphone to be equipped with AI. After an update, it will be able to help produce quick responses for us in WhatsApp, by consulting the message history, and suggesting content accordingly. Gemini will also be able to quickly summarize voice notes or transcribe them. “This is one more step towards our vision: to bring you the best AI collaborator in the world,” said Sissie Hsiao, Google vice president of Bard, Google’s conversational AI tool, on Wednesday. We will have to wait until 2024 for the deployment of new features, such as advanced help in solving math problems. But, according to Sundar Pichai, Google has already entered the future with both feet. “This new era of models represents one of the greatest scientific and engineering efforts we have undertaken as a society,” he said in a statement.