Threads will launch an AI bot capable of conducting conversations and fact‑checking, like Musk’s.

Meta AI launches a public chatbot on Threads

Meta AI has begun testing a new feature on the Threads platform that allows the Meta AI bot to reply directly under posts and add additional context—similar to how Grok works on X. The official account @meta.ai will appear in users’ feeds as part of an early beta.

Where the feature will be available
According to Engadget, the bot will first launch in:

- Malaysia
- Saudi Arabia
- Mexico
- Argentina
- Singapore

Users can ask questions, and Meta AI posts answers in comments. At any time a discussion participant can invite the AI to their reply to verify facts or clarify information.

Notification management
It is possible to completely disable notifications from @meta.ai—both alerts and replies. This is done to reduce risks of AI errors and abuse by hackers. Similar problems already occurred with Grok on X: the bot sometimes generated incorrect statements about historical figures.

Current context of the update
Testing is part of a large-scale Meta AI update, now running on the Muse Spark model. At the same time, the company is experimenting with chats in WhatsApp and other apps where users can privately request discussion context from the bot without exposing the query to others.

New device capabilities
The Spark model is also used in the updated Live AI version—a feature that was previously exclusive to smart glasses. With it, you can get real‑time answers about what the smartphone camera “sees.” Soon support for Spark will expand to Ray‑Ban and Oakley in the U.S. and Canada, and by summer to display-equipped frames.

Mark Zuckerberg also mentioned plans to create AI agents at the OpenClaw level, running on Muse Spark.