Google is developing a new toggle that will allow users to disable the Guided Vision feature in Gemini Live, giving people more control over when the AI can access their device's camera. The feature was spotted in an unreleased version of the Google app for Android and represents a significant shift toward user privacy in vision language models (VLMs), which are AI systems that can both read text and analyze images.

What Is Guided Vision in Gemini Live?

Guided Vision is the camera-based feature that lets Gemini Live "see" what's around you in real time. When you point your phone at something, the AI can identify objects, read text, understand scenes, and offer immediate suggestions or answers. Google first launched this capability alongside the Pixel 10 and later rolled it out to other Android and iOS devices. The feature combines real-time camera input with Gemini's conversational abilities, making the AI assistant more helpful when you're out in the world.

The practical applications are broad. You can use Guided Vision while shopping to compare products or get styling advice, point your camera at unfamiliar tools to learn what they are, or even get feedback on photos and social media content by sharing your screen with the AI. The technology preserves the conversational flow, allowing you to ask follow-up questions naturally.

Why Is Google Adding a Privacy Control Now?

The new toggle addresses a growing concern among users: the idea that apps have constant access to their camera. By letting people decide when Gemini can see through their camera, Google is giving users more control over their personal information and surroundings. This move reflects broader industry recognition that powerful vision capabilities need to come with meaningful privacy safeguards.

The toggle was discovered in Android app version 17.38.5.sa.arm64, appearing as a setting called "Guided Vision in Live" within the Gemini options. Screenshots show users will be able to switch the feature on or off as they prefer. However, Google has not yet officially announced when this setting will launch or provided detailed information about what happens when users disable it.

How to Manage Your Camera Privacy in Gemini Live

• Disable Guided Vision: Once the toggle rolls out, you'll be able to turn off camera-based AI assistance entirely, preventing Gemini Live from accessing your camera feed during conversations.
• Selective Camera Use: You can choose to enable Guided Vision only when you need it, rather than leaving it available all the time, giving you granular control over when the AI can see your surroundings.
• Monitor App Permissions: Even with the toggle in place, check your device's camera permissions regularly to ensure only trusted apps have access to your camera.

What's Next for Vision Language Models and Privacy?

This privacy control is part of a larger wave of AI updates from Google. The company recently unveiled Gemini 3.5, which includes Live Translate for natural speech-to-speech translation across 70 languages, and Gemma 4 12B, a model that runs locally on consumer devices with just 16GB of memory. These developments show Google embedding AI more deeply across its ecosystem while also recognizing the need for user control.

The timing is significant. As vision language models become more capable and integrated into everyday tools, the tension between functionality and privacy grows sharper. Google's decision to add a privacy toggle suggests the company is listening to user concerns about camera access, even as it expands what these models can do. The feature is expected to roll out in a future Android app update, though no official launch date has been announced.

For users who rely on Gemini Live for productivity and convenience, this toggle offers peace of mind. For those who prefer to keep their camera private by default, it removes a potential barrier to using the AI assistant. Either way, the move signals that privacy controls are becoming table stakes for vision language models in consumer applications.