Our Take
Meta is copying X's playbook but with better guardrails than Grok's Hitler-praising mishaps.
Why it matters
Social platforms are racing to keep users from leaving to search elsewhere for context. Meta's safety advantage could determine whether AI-assisted discourse stays productive or devolves into misinformation amplification.
Do this week
Social media managers: Test the @meta.ai mention feature in beta countries to understand response quality before broader rollout.
Meta AI joins Threads conversations in five countries
Threads is testing a Meta AI integration that mirrors X's Grok functionality. Users with public accounts can mention Meta AI in posts or replies to get contextual information about trending topics and breaking news (per Meta statement to TechCrunch).
The feature is currently in beta testing in Malaysia, Saudi Arabia, Mexico, Argentina, and Singapore. When summoned, Meta AI responds as a public reply from the @meta.ai account, matching the language of the original post.
Users can ask questions like "why are people talking about the World Cup this month?" or "whose Met Gala looks are trending right now?" Meta AI processes these mentions and provides public responses within the conversation thread.
Platform retention trumps search redirection
The integration positions Threads as a self-contained information hub rather than a jumping-off point to external search. This directly mimics X's strategy with Grok, where users frequently ask "is this real?" or "explain this" to stay within the platform ecosystem.
Meta acknowledges the risks, referencing Grok's problematic outputs including posts praising Hitler. The company claims stronger safeguards than Grok but admits the potential for similar issues remains untested at scale.
Users can mute @meta.ai, mark responses as "Not interested," or hide specific Meta AI replies. The company plans to refine the experience based on beta feedback before broader expansion.
Content moderation meets conversational AI
The beta's geographic limitation to five countries suggests Meta is stress-testing both technical performance and content safety in diverse linguistic and cultural contexts. This approach contrasts with Grok's broader but more chaotic deployment.
For platforms considering similar integrations, Meta's phased rollout provides a template: start with limited geography, maintain user control over AI visibility, and prepare content filtering specifically for conversational AI outputs that appear in public feeds.
The success metric will be whether Meta AI can provide useful context without generating the viral misinformation incidents that have plagued Grok on X.