Google stands on the verge of a major step forward in AI. Leaks from the past 48 hours reveal an experimental feature called Gemini Spark. It promises to move the company’s chatbot beyond quick answers. Instead it would run in the background. Handling email triage, meeting prep and custom news updates without constant user prompts.
The details surfaced first through an onboarding screen inside the Gemini web app. Then beta code in the Google app version 17.23 confirmed the name. Android Authority reported the agent could clean junk from Gmail inboxes or assemble notes before key meetings. It might generate a tailored news digest that follows stories as they develop. These tasks go far past what current Gemini versions deliver.
But here’s the catch. Spark needs deep access to personal data. According to the leaked welcome text it pulls information from connected apps, user skills, chat histories, scheduled tasks, logged-in websites, personal intelligence signals, location data and more. It can share that information with third parties. Names. Contact details. Files. Preferences. Even data users might consider sensitive. The disclosures appear in screenshots shared by developer Fandu on X and analyzed across tech sites.
TestingCatalog captured the full onboarding message. “Let Gemini do more as your everyday AI agent, ready 24/7 to help with your inbox, online tasks, and more.” The screen continues. “The more you use Gemini Spark, the better it understands you and what you want to accomplish.” That adaptive quality sits at the heart of the pitch. TestingCatalog noted the feature builds on an earlier internal codename called Remy. It also ties to an existing Gemini Agent option limited to Google AI Ultra subscribers.
Privacy warnings come attached. The text states Spark is experimental. It is designed to ask permission before sensitive actions. Yet it may share information or complete purchases without asking. Users must supervise it. The notice explicitly advises against relying on the agent for medical, legal, financial or other professional advice. Review the risks, it says. Such language reflects the tension Google faces. Deliver genuine utility. Manage expectations around autonomy and errors.
9to5Google dug into the Android app beta. The publication found Spark appearing in a redesigned navigation drawer. A two-tab layout splits the experience between Chat and Agent. Users create new tasks in the Agent view. They see lists of active work and items scheduled for specific times. The icon itself looks like the Gemini spark with added momentum. A comet streaking across the interface. 9to5Google reported the name replaced earlier references to a generic Gemini Agent in recent beta builds. This marks the second consecutive week of dual beta releases from the company. Activity has clearly accelerated.
Capabilities described in the leaks paint a picture of practical productivity gains. Declutter an inbox by summarizing newsletters and offering unsubscribe options. Prepare concise briefs with relevant documents before meetings. Track news stories in depth and monitor how they evolve over days or weeks. The agent can create reusable skills for recurring tasks. Feed variables through prompts to adjust outputs without rewriting instructions each time. Android Authority highlighted potential to summon Google Workspace tools, control Chrome browser sessions and index data across multiple applications at once.
Some limitations remain visible even in the early leaks. Current screenshots show integration mainly with Google’s own apps. Third-party support appears possible but unconfirmed in the visible code. The agent stops short of full computer control. It cannot yet match projects like Anthropic’s Claude Computer Use or OpenAI’s Operator in raw system-level actions. Still the direction feels clear. Google wants agents that operate independently once given goals.
And the timing adds weight. Google I/O opens next week on May 19. Multiple outlets suggest Spark could feature prominently in the keynote. The leaks arrive just days before the event. A pattern familiar from past product rollouts. Earlier this week the company previewed a refreshed Gemini interface along with basic automation tools. Spark represents the more ambitious leap. From reactive chatbot to proactive operator embedded across the product lineup.
Recent coverage adds fresh context. Android Headlines noted the agent may manage emails, build custom skills and function without constant human oversight. One report from the site published hours ago indicated an official reveal could come during the I/O keynote. Android Headlines emphasized the always-on nature. The agent learns from behavior. It connects apps and executes multi-step processes in the background.
Concerns about data sharing have already surfaced in online discussions. Reddit threads linked to the 9to5Google story show users reacting to the list of data sources and the explicit warning about purchases without confirmation. Some expressed discomfort at the breadth of access. Others see the trade-off as worthwhile if the agent delivers on promised efficiency. Google has provided settings to clear remote browser data including logins and execution states. Users can disable connected apps and personal intelligence features. Activity controls in Gemini Apps Activity offer another layer of management.
The emergence of related model names fuels speculation about underlying technology. Leaks mention Spark Robin described as offering rich visual responses. A new MCP Tool Testing category appeared in the model selector. These elements hint at expanded multimodal abilities and tighter integration with external tools. Whether Spark relies on a dedicated model or draws from the broader Gemini family stays unclear from available code strings.
Google has not commented publicly on the leaks. That silence is standard before major events. Yet the volume of coordinated details across web app, Android beta and third-party testing suggests internal testing has reached an advanced stage. Power users who already pay for premium Gemini access stand to benefit first. Inbox management alone could reclaim hours each week. Meeting preparation that surfaces the right files without manual search offers immediate value. Custom news feeds that evolve with events could change how professionals stay informed.
Of course execution will decide success. Early agent experiments from competitors have shown the gap between promised autonomy and real-world reliability. Hallucinations, incorrect actions and unexpected data leaks remain risks. The disclaimers in Spark’s own onboarding text acknowledge those realities. Supervise the agent. Don’t treat it as infallible. Those words may prove as important as any feature list.
Industry watchers will dissect the I/O presentation for clues about broader strategy. How deeply will Spark integrate with Android, Workspace and Chrome? Will it expand beyond Google services? Can the company balance powerful capabilities with the transparency users demand? The leaks provide a preview. They do not deliver final answers.
What feels certain is the trajectory. Chat interfaces alone no longer satisfy. Users want systems that act. That monitor, prioritize and execute on their behalf. Gemini Spark, if delivered as described, positions Google to compete directly with agent-focused offerings from Anthropic, OpenAI and others. The race has moved from who answers questions fastest to who finishes the work while users focus elsewhere.
Developers and enterprise teams may find particular interest in the skills framework. Reusable templates for common workflows could accelerate adoption inside organizations. A marketing team might build a skill that gathers campaign metrics, drafts reports and schedules reviews. An operations group could automate vendor communications and invoice tracking. The variable-driven prompts add flexibility without requiring constant reprogramming.
Privacy advocates will scrutinize the data flows. The ability to share sensitive information with third parties to complete tasks introduces new vectors for exposure. Google will likely need to detail exactly which partners receive data and under what conditions. Clear audit logs and granular controls could help address concerns. Without them trust may prove harder to earn than technical capability.
So the leaks offer more than product teases. They signal a philosophical change in how Google views AI assistance. Less oracle. More colleague. One that works alongside users. Learns preferences over time. Takes initiative within defined boundaries. Whether that vision holds up under real usage will emerge in the months after I/O. For now the screenshots and beta strings have opened the window. What Google shows on stage next week will determine how wide that window becomes.
WebProNews is an iEntry Publication