Google names upcoming Gemini AI agent 'Spark,' adds autonomous task execution to mobile app
Google is preparing to launch Gemini Spark, an autonomous AI agent that will operate within the Gemini mobile app. According to code found in Google app beta version 17.23, Spark can access connected apps, personal data, and location to execute tasks like managing inboxes and scheduling meetings, though Google warns it may occasionally act without permission.
Google names upcoming Gemini AI agent 'Spark,' adds autonomous task execution to mobile app
Google is preparing to launch Gemini Spark, an autonomous AI agent that will operate within the Gemini mobile app, according to code discovered in Google app beta version 17.23 released today.
The agent, previously referred to internally as "Gemini Agent," will appear as a separate section in the Gemini app's navigation drawer with a two-tab layout split between "Chat" and "Agent." Users will be able to create, schedule, and monitor active tasks through a dedicated interface.
Data access and capabilities
Gemini Spark will access user information from connected apps, chat history, tasks, logged-in websites, Personal Intelligence data, and location data. According to disclosures in the app code, the agent "can share necessary info with third parties" to complete tasks, including names, contact information, files, preferences, and potentially sensitive information.
Confirmed capabilities include:
- Email management: summarizing newsletters, archiving messages, and unsubscribing from email lists
- Meeting preparation: generating briefs with relevant information before scheduled meetings
- News curation: creating custom news digests and tracking story developments over time
Experimental status and warnings
Google labels Gemini Spark as "experimental" in the code strings. The company warns that "while it is designed to ask for your permission before taking sensitive actions, it may do things like share your info or make purchases without asking." Users are advised to supervise the agent and avoid relying on it for medical, legal, financial, or professional advice.
The feature is already appearing for some beta testers ahead of Google I/O 2026, though no official release date has been announced.
What this means
Gemini Spark represents Google's entry into autonomous AI agents that can execute multi-step tasks across apps without continuous user input. The disclosure that it may occasionally act without permission highlights the control challenges facing agent systems as they gain more autonomy. The timing ahead of I/O 2026 suggests Google plans a formal announcement at its developer conference, likely competing with similar agent features from OpenAI and Anthropic.
Related Articles
Google rolls out computer use in the Gemini app for Mac for multi-step desktop tasks
Google announced on November 19, 2026 that computer use is rolling out in the Gemini app for Mac. Gemini can carry out multi-step work across local files, desktop applications and websites, working in an isolated background window with a live preview in the chat.
Google builds isolated 'Mirror Dimension' web to train and test Chrome AI agents
Google's Chrome team built the Mirror Dimension, an isolated web ecosystem it fully controls, to train and test Gemini-based browsing agents without exposing them to the open internet. Google is detailing the project in a blog post, according to Wired.
Tunix agentic GRPO lifts Gemma 4 E2B from 9.4% to 95% win rate in Plane Strike
Google's developer blog details how Tunix's agentic GRPO pipeline trained Gemma 4 E2B to play Plane Strike with tool calls. The win rate on held-out boards rose from 9.4% to 95% after 100 steps on a single Cloud TPU v6e-8 host.
Google previews Tool Retrieval in Gemini Interactions API to load tool schemas only when needed
Google announced Tool Retrieval on November 19, 2026, available in preview in the Gemini Interactions API. Agents can keep large tool libraries on standby and load full parameter schemas only when the current task needs them.
Comments
Loading...