According to Beating monitoring and developer @mrfanduuuuu’s screenshot, Google is testing Gemini Spark, a 24-hour autonomous AI Agent on Gemini’s web version, with official announcement expected on May 19 at Google I/O. Spark executes multi-step tasks including inbox management, form filling, and shopping orders by extracting context from connected Google apps, chat history, location data, and Personal Intelligence.
However, Gemini’s risk warning states the agent may share user information or complete purchases without confirmation, despite being designed to request permission for sensitive operations. The system also stores remote browsing data including login credentials. According to TestingCatalog, Spark is an upgrade from an internal agent codenamed “Remy”, with Google positioning it directly against Anthropic’s Conway agent and OpenAI’s agent platform.
Related News
WhatsApp introduces Meta AI “disappearing” chat, prompting concerns over accountability mechanisms as messages automatically vanish
Mistral AI in talks with European banks to develop Mythos as a replacement for internet security models
Google plans to partner with SpaceX, laying out its “Capturing the Sun” program for space AI computing