Google is stepping up its AI game with Mariner, a prototype AI agent built on its newly upgraded Gemini 2.0 framework. Mariner is designed to handle web-based tasks autonomously, like research, booking, and shopping — in other words, to become your online personal assistant, Forbes reports.
(Tap or click the headline above to read this story with all of the links to external sources.)
The brain behind Mariner: Mariner capitalizes on the enhanced reasoning, planning and multi-step capabilities of Gemini 2.0. Unlike its predecessor, which focused on integrating text and image processing, Gemini 2.0 introduces features like Deep Research — which compiles detailed reports mimicking a human assistant. Mariner can fill your online shopping cart or analyze complex datasets, but the final decisions are still up to you.
How does Mariner stack up in the AI race? While Google juggles between innovation and legal issues, OpenAI is continuing to enhance its GPT-4 model, integrating it into ChatGPT while securing enterprise partnerships. Microsoft is refining its AI powered Copilot, and Amazon’s AWS is tailoring Bedrock to meet company demands.
Mariner is just one agent in Google’s arsenal: Google has also previewed tools like Jules for coding assistance and extensions for decision making in video games — showing dedication into providing tools for both easy workflow and encouraging developers.
But there’s a catch: It’s not foolproof yet, Project Manager Jaclyn Konzelmann stated. The technology is still in its experimental phase, which is why it’s only receiving a limited release, but they hope to make it available to the public by 2025.