Loading article…
Google unveils Gemini Spark, an always‑on AI agent that scans emails, calendar and files. The rollout to AI Plus, Pro and Ultra users hinges on users opting‑in
Google unveiled Gemini Spark, an always‑on AI assistant that can draft emails, generate to‑do lists and scan credit‑card statements, but it only works if users opt‑in to give the system continuous access to Gmail, Calendar, Drive and even local Mac files [1].
| At a glance | |
|---|---|
| Product | Gemini Spark AI agent |
| Launch | I/O 2026 announcement |
| Availability | Google AI Plus, Pro, Ultra subscribers |
| Data access | Requires opt‑in to Gmail, Calendar, Drive, Photos, local files |
The I/O 2026 keynote introduced Gemini Spark alongside Daily Brief, which pulls updates from Gmail and flags Calendar events for subscribers of Google AI Plus, Pro and Ultra [1]. Google also expanded its AI‑enhanced Gmail inbox, now able to generate custom to‑do lists and draft replies based on email content [1]. These features build on the 2024 rollout of Gemini across Workspace apps—Gmail, Docs, Sheets, Slides and Drive—where the chatbot could sift through files and draft messages [1].
Google’s approach contrasts with Apple’s device‑first AI strategy that keeps personal data on‑device, limiting the breadth of assistance but avoiding the “always‑on” data collection model [2]. OpenAI and Microsoft are pursuing similar agent visions, but they typically require users to connect external apps rather than granting a single platform deep, cross‑service access [2]. By positioning Gemini Spark as a 24/7 personal assistant that can also plug into third‑party services such as Canva, OpenTable and Spotify, Google is betting that convenience will outweigh privacy concerns [1].
Google claims “millions of people” are already using its “Personal Intelligence” feature, which lets Gemini reason across Gmail, Photos, Search and YouTube history without prompting [1]. If adoption scales, the company could lock users deeper into its ecosystem, reinforcing its AI revenue stream. However, the reliance on extensive personal data revives scrutiny of Google’s privacy track record, a factor that analysts say could influence whether users stay with Google or shift to competitors with tighter data controls [2].
Google’s AI future now hinges on whether users will trade broad personal data access for the promised productivity gains, a decision that will shape the competitive landscape of consumer AI assistants.
Coverage is mostly measured — 157 of 169 reports stay neutral.
Every Monday — the token unlocks, Fed dates & catalysts set to move crypto and markets this week. So you’re never blindsided.
Free · 3-min read · one-click unsubscribe
AI-assisted synthesis by the TrendWatcher Editorial Desk · sourced from 4 outlets · Jun 17, 2026 · How we report
Gemma 4 E2B is a lightweight AI model optimized for the Tensor Processing Unit in Pixel devices, allowing AI tasks to be processed locally for faster performance and offline use.
The institute reported that AI Overview and AI Mode missed suicide cues, mischaracterized eating disorder symptoms, and gave step‑by‑step instructions for creating deepfakes, with no parental controls to disable these features.
Google said the report tested a narrow set of queries that do not reflect typical usage, could not reproduce many of the highlighted responses, and emphasized existing safety guardrails and the ability to turn off Search entirely for child accounts.