Gemini 2.0 launch: What are AI Agents and the types Google is working on

Google DeepMind, the company’s AI team, has announced a series of experimental projects that showcase the potential of AI agents to assist humans in various tasks, from navigating the web to coding and even playing video games.
announced Project Mariner, an AI Agent for the Web; Jules, an AI coding companion and AI Agents for gaming.
An AI agent is a digital assistant that can not only understand a user’s requests but also take actions to fulfill them, learning and adapting along the way. Consider an AI agent as Tony Stark’s JARVIS, an AI program that helped him to innovate.

Project Mariner is an early research prototype that explores the future of human-agent interaction within a web browser. Powered by Gemini 2.0, Mariner can understand and reason about information on a web page, including text, images, code, and forms. Google says that it can then perform tasks like filling out forms and navigating websites.
“For example, Project Mariner can only type, scroll or click in the active tab on your browser and it asks users for final confirmation before taking certain sensitive actions, like purchasing something,” the company said.

CEO Sundar Pichai said that people in Google are now using AI for coding. Jules is an experimental AI-powered code agent designed to assist developers within their GitHub workflow. It can analyse issues, devise solutions and execute code under a developer's guidance.

Gaming is seen as one of the biggest use cases of AI and AI agents. Google DeepMind is also exploring how AI agents can enhance gaming experiences. These agents can analyse gameplay, offer suggestions, and even access Google Search to provide relevant information in real-time.
“We're collaborating with leading game developers like Supercell to explore how these agents work, testing their ability to interpret rules and challenges across a diverse range of games, from strategy titles like “Clash of Clans” to farming simulators like “Hay Day”,” Google said.
This comes a week after Google introduced Genie 2, an AI model that can create a variety of playable 3D worlds from a single image.
Beyond gaming, Google DeepMind is experimenting with AI agents for robotics, leveraging Gemini 2.0's spatial reasoning capabilities to enable assistance in the physical world.