Google's Gemini Spark planned a birthday party but missed the most important person

Google has unveiled Gemini Spark, a new AI agent that can access your personal information. Using birthday planning as a test case, a Wired journalist found that the technology already works fairly well. She also found strange blind spots in how it perceives her life.

How Spark works

Google Gemini Spark is a new-generation AI agent that can work not only with text but also with a user's personal content. It is not just another chatbot with a fashionable interface. Spark can access highly private data and use it to help with tasks.

Technically, the agent can:

• Read and analyze emails from Gmail
• Browse documents in Google Docs and Google Drive
• Check events in Google Calendar
• Synthesize information from different sources
• Suggest actions based on the patterns it identifies

The logic is straightforward: if you give an AI full access to your life, it can help far more than an ordinary chatbot that knows nothing about you.

The birthday test

The Wired journalist decided to try Spark in practice. The task was simple: plan the perfect birthday. She gave the agent full access to her email, documents and calendar, everything Google knows about her life.

The result was decent. Spark pulled details about her food preferences from her emails, identified potential guests and checked their schedules in shared calendars. The agent even proposed several perfectly reasonable party options, taking into account the season and venue availability. Technically, it is impressive.

But there is one strange problem

Here things get interesting. Despite full access to her correspondence and documents, Spark did not notice the most important person in the journalist's life: her boyfriend. He appeared in her emails, was mentioned in her documents and was listed in shared calendar events. Yet when the agent analyzed all the information, it somehow skipped this obvious detail.

This is not just an interface glitch or some bug. It points to a fundamental problem: even with access to your personal life, an AI can still fail to grasp what matters and what does not. The agent can process data, but its reasoning remains shallow.

What this means

Spark shows that AI agents have become advanced enough for serious tasks involving personal data. But the birthday story is a reminder that full access to information does not guarantee a correct understanding of context. Before giving an AI full access to your email and calendar, it is worth understanding its limitations.