xAI tests screen sharing on Voice Mode for Grok web

xAI continues to develop the voice mode for Grok on the web, now introducing a screen-sharing option. This feature is accessed via a new camera button, prompting users to share their screen with Grok. While Google AI Studio has a comparable capability, other major competitors like ChatGPT and Google’s web interface currently do not. For now, Grok’s implementation appears to be in an early stage; Screen content isn’t yet processed by the model, meaning the screen sharing function doesn’t allow Grok to “see” or interact with what’s displayed. This points to a development phase where additional iterations are likely before it becomes a fully working part of the tool.

In addition to the screen sharing update, new references to tool integrations have surfaced within Grok’s internal configuration. Two new capabilities are being tested:

1. Keyword search on X
2. Semantic search on X

Presently, Grok can already access X data through the API, but these additions suggest a move towards more refined, possibly context-aware search operations. As Grok’s toolset is typically hidden from end-users, the choice of which tool is used will likely be handled automatically by the model itself rather than exposed for manual selection.

These changes, if implemented, could benefit users who need richer, more interactive support for browsing or researching information on X. The addition of both screen sharing and advanced X search would give Grok unique capabilities within the AI assistant space, especially for those who rely on live web content or social media data in their workflow.

However, the long-anticipated release of Grok 3.5 remains on hold, with no official timeline shared by xAI, leaving many users waiting for the next major update. This ongoing development fits xAI’s current strategy of gradually layering new features into Grok, with a clear focus on deepening integration with X’s ecosystem and experimenting with novel ways of interacting via the web.