Google showed a Jeanent that helps to pass the games, watching the actions of the player

Google has introduced a new version of its artificial intelligence platform Gemini 2.0, which can be used to create AI agents for advice and tips in video games. Agents are able to analyze gameplay and offer optimal strategies, and users will be able to receive recommendations in real time.

Image source: Supercell, theverge.com

As Google DeepMind CEO Demis Hassabis and CTO Koray Kavukcuoglu wrote in a company blog, the agents “make inferences about the game based solely on on-screen actions and make decisions in real time.” They can also use Google knowledge searches to provide more comprehensive information. Agents are already being tested in popular games such as Clash of Clans and Hay Day from Supercell. However, as The Verge writes, the technology is in the early stages of development, and it is not yet clear how useful such mentors will be for players.

In addition to helping with video games, Google is developing other uses for Gemini 2.0. For example, work is underway on the Genie 2 project, which creates virtual game worlds based on just one image. While these worlds only remain stable for about a minute, experts see great potential for the technology.

The company also introduced the multimodal capabilities of Gemini 2.0 Flash, which are already available to developers through the API in Google AI Studio and Vertex AI. The new version features fast performance, improved data processing, and the ability to generate images, text, and text-to-speech in multiple languages. These functions are already being tested by early partners, and in January the platform will be available to a wider audience.

Other projects include Project Astra, Mariner and Jules. Designed for use on Android devices, Project Astra has been enhanced with Gemini 2.0. Now the Astra-powered assistant can speak multiple languages, use Google Search, Maps and Lens, and remember more data while maintaining privacy. Project Mariner explores how AI can assist in the browser by recognizing text, images, code and other interface elements, while Jules is designed to help developers with their GitHub workflows.

AI agents can also be used in the physical world. Google research shows that Gemini 2.0 is capable of using spatial reasoning in robotics. Although these developments are at an early stage, the company itself sees great potential in creating agents that are able to interact with the real environment.

To minimize risks and ensure the safety of its AI services, the company conducts extensive testing and collaborates with trusted testers and external experts. In the near future, Google plans to integrate Gemini 2.0 capabilities across all of its products, including the Gemini mobile app.