Google has unveiled a new prototype called Mariner, an artificial intelligence agent that can use websites and perform actions on behalf of users. Based on the company’s Gemini 2.0 technology, Mariner is designed to work with a human in the loop, ensuring tasks are completed accurately.
According to Jaclyn Konzelmann, a Google project manager, Mariner enables users to type requests into their web browser and have it take actions on their behalf. The AI system learns from vast amounts of data, including text, images, and sounds, allowing it to recognize patterns and make informed decisions.
Mariner’s capabilities include filling virtual shopping carts, completing tasks online, and interacting with websites in a human-like manner. However, the technology is still experimental and prone to mistakes. Demis Hassabis, who oversees Google’s core AI lab, emphasizes that Mariner is designed to work alongside humans, not replace them entirely.
Google plans to share Mariner with a small group of testers outside the company but has not announced a wider release yet. The AI agent is also part of a larger effort to develop a universal assistant, as envisioned by Sundar Pichai, Google’s chief executive.
Source: https://www.nytimes.com/2024/12/11/technology/google-ai-agent-gemini.html