Threads is testing what happens when AI becomes part of the conversation, not just something sitting in a separate chat box.
According to TechCrunch, Meta is testing a new Threads integration that lets users mention Meta AI in posts or replies to get more context, recommendations, or explanations inside public conversations.
The test is currently live in beta in Malaysia, Saudi Arabia, Mexico, Argentina, and Singapore. Users with public accounts can tag Meta AI with questions about trends, news, sports, entertainment, or recommendations. Meta AI then replies publicly from the @meta.ai account, using the language of the original post.
If that sounds familiar, it should. The behavior is similar to how people use Grok on X: asking whether something is real, requesting context, or turning an unfolding conversation into a quick explainer.
Threads wants context to stay inside the feed
This is not just an AI feature. It is a retention feature.
Threads has been trying to become a real-time conversation layer, especially around news, culture, and live events. Adding Meta AI gives the platform a way to answer questions without pushing users back to search, Google, ChatGPT, or X.
It also changes the shape of public discussion. Instead of users leaving the thread to look something up, the answer can arrive directly inside the conversation, visible to everyone else.
The risk is visibility
Public AI replies are powerful because they are ambient. They are also risky for the same reason.
X has already shown what can happen when an AI account becomes a visible participant in public discourse. Meta says users will be able to mute @meta.ai, hide replies, or mark them as not interested, and the company will use early feedback before expanding the test.
Still, the direction is clear. Social platforms do not want AI to live in a sidebar. They want it embedded in the social fabric, answering questions, shaping context, and keeping attention inside the app.
Threads may not want to be X. But it clearly wants the part of X where the feed can explain itself.