Google is shifting towards using AI agents to bring users information from the web, moving away from traditional Search.
At I/O 2025, Google introduced AI agents that allow users to ask questions, shop, and conduct research in a more personalized manner.
The new AI features include conversing with AI agents, Project Mariner for hands-off tasks, Deep Research for generating reports, and Project Astra for multimodal experiences.
AI agents dominated the discussion at I/O 2025, signaling a significant shift in Google's approach to Search and the web.
Google aims to replace traditional Search with AI agents that can fetch and present information in various ways tailored to users' preferences.
The introduction of Gemini models supporting Anthropic's MCP standard signals Google's commitment to AI and agent-based technologies.
Microsoft also envisions an 'open agentic web,' emphasizing the importance of connecting agents and data sources for a seamless online experience.
However, the use of AI agents raises concerns about the impact on internet economy, particularly for publishers relying on ad revenue and content creation.
The potential for AI systems to generate misinformation or hallucinations poses challenges for trust in online information and consistency in AI models.
Google's push towards AI agents may revolutionize the web, but it also presents risks and challenges that need to be addressed for a more reliable and trustworthy online ecosystem.