Threads has begun testing a Meta AI mention feature that lets public-account users tag @meta.ai inside any post or reply to pull in real-time context, recommendations, and explanations. The beta is live in Malaysia, Saudi Arabia, Mexico, Argentina, and Singapore, and the design closely tracks how Grok functions on X. Meta confirmed the rollout in an email, framing it as a way to surface breaking news and trend context without leaving the app.
The invocation pattern is simple. A Threads user writes something like "Why are people talking about the World Cup this month?", "Whose Met Gala looks are trending right now?", or "How are the Knicks doing in the playoffs?", tags Meta AI, and the system posts a public reply from the @meta.ai account. Meta AI responds in the same language as the original post.
The product strategy is plain. Threads has spent the last two years positioning itself as a conversation layer for news and culture, and an in-thread assistant turns the feed itself into a query interface. Instead of bouncing to a chatbot tab or a search engine, the user gets the answer where the conversation is already happening — and so does everyone else reading the thread.
Key facts
- 01Threads is testing a Meta AI mention feature that lets public-account users tag @meta.ai for context inside posts and replies.
- 02The beta is live in five markets: Malaysia, Saudi Arabia, Mexico, Argentina, and Singapore.
- 03Meta AI replies are posted publicly by the @meta.ai account and respond in the language of the original post.
- 04Users can mute @meta.ai, mark replies as 'Not interested,' or hide Meta AI replies under their own posts.
- 05The feature mirrors Grok's role on X, where users routinely ask the chatbot to verify or explain posts.
That last part is the lever. Public Meta AI replies become discoverable content in their own right, the same dynamic that has made Grok a fixture under viral X posts, where users routinely reply "Is this real?" or "Explain this." The chatbot answer becomes part of the social object, not a private side conversation.
“Threads is positioning itself as a place where users ask Meta AI to explain a trend mid-thread, the same behavior pattern that turned Grok into a fixture on X.”— Jaeden Schafer
Meta is layering controls on top. Users can mute @meta.ai outright, tap "Not interested" on any Meta AI post to see fewer of them, or hide a Meta AI reply that appears directly on one of their own posts. The company said it plans to refine the experience based on early feedback before widening the test.
Choosing Malaysia, Saudi Arabia, Mexico, Argentina, and Singapore as launch markets is notable. None of them are the US or Western Europe, where regulatory scrutiny of AI-generated content on social platforms is highest. The pattern mirrors how Meta has historically piloted features in growth markets first and exported them once the kinks are worked out.
The Grok comparison cuts both ways. xAI's chatbot demonstrated the appetite for in-feed AI answers, but it also demonstrated the hazard — Grok at one point generated posts praising Hitler, a failure mode that drew regulator and advertiser attention. An assistant that posts publicly under a verified-looking handle inherits the platform's reputation every time it answers.
Meta said its safeguards are stronger than what shipped with Grok, though it stopped short of detailing them. The real test will be whether Meta AI holds up to the same adversarial prompting that has produced embarrassing outputs across nearly every major consumer chatbot, and how quickly Meta can pull or correct replies that go wrong in a public thread.
There is also a competitive read here for the broader social-plus-AI category. X built Grok as a differentiator. Meta is responding with an assistant tied to a much larger user base across Threads, Instagram, WhatsApp, and Facebook, and the @meta.ai account effectively gives the company a persistent first-party presence inside conversations on Threads.
Skeptics will point to the obvious risks: hallucinated answers reaching audiences before corrections, AI replies crowding out human conversation, and the awkward fact that no one has yet figured out a clean moderation model for a chatbot that posts publicly at scale. Meta has not said how it will handle Meta AI replies that go viral with errors, or whether corrections will be issued as new replies or edits.
For Meta, the prize is engagement loops that don't currently exist on Threads — users hanging around longer because the answer to their question is one mention away, and Meta AI generating its own follow-on conversations. For the rest of the AI industry, this is another data point that consumer AI distribution is shifting from standalone apps to features bolted into the platforms people already use, and the companies that own the surface area have an advantage that no model benchmark will close.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
