Google has officially entered the "agentic era" of artificial intelligence, announcing a new research prototype called Project Mariner and providing updates to its existing Project Astra. The tech giant's latest advancements in AI technology are poised to revolutionize the way we interact with virtual assistants and complete tasks online.

Project Astra, a virtual assistant first demoed at Google I/O earlier this year, has expanded its testing program to more users. The AI "universal agent" can process text, images, video, and audio in real-time and respond to questions about them. Google has also integrated Astra into several of its products, including Search, Lens, and Maps. The updates aim to improve Astra's understanding of various accents and uncommon words, increase its in-session memory to up to 10 minutes, and reduce latency.

Project Mariner, a previously unannounced experiment, is an AI agent that can take control of a user's browser using a Chrome extension to complete tasks. Although still in its early stages, Mariner has entered testing with a pool of "trusted testers." In a demo, Google showcased Mariner's ability to add items to a user's online shopping cart, albeit slowly. The company acknowledges that speed is an area that needs improvement, but the technology has the potential to simplify online tasks.

Google's CEO, Sundar Pichai, defines agents as models that "can understand more about the world around you, think multiple steps ahead, and take action on your behalf, with your supervision." However, the release of these agents to the general public is still uncertain due to concerns about their unpredictability and vulnerability to hijacking via prompt injections. Google has stated its intention to protect against such attacks by prioritizing legitimate user instructions.

The "agentic era" marks a significant shift in AI development, with companies like OpenAI and Anthropic also working on similar technologies. While these agents are not yet ready for mass-market scale or access to sensitive data, they have the potential to revolutionize the way we interact with technology. Google's latest updates demonstrate its commitment to pushing the boundaries of AI research and development.

Despite the imperfect nature of these prototypes, they are still impressive to see in action. As Google continues to refine its agents, we can expect to see significant advancements in the field of AI. The question remains, however, when these systems will be ready for public release and what they will look like when they finally arrive.

In the meantime, Google's "agentic era" announcement serves as a reminder of the rapid progress being made in AI research and development. As the technology continues to evolve, we can expect to see new and innovative applications of AI in various industries and aspects of our lives.