Meta announced late Tuesday that it is testing the integration of a Meta AI account into Threads — an AI assistant disguised as a real user. Users can tag @MetaAI on posts to get real-time answers, such as "Why is everyone talking about the Tamkang Bridge lately?" or "How much did Taipower spend on its logo redesign?" The test is currently limited to Argentina, Malaysia, Mexico, Saudi Arabia, and Singapore; Taiwan is not included.
How @MetaAI works
Unlike traditional AI helpers hidden in search bars or side panels, Meta AI appears as a standalone account. When tagged in a thread, it replies directly, occupying visual space in the comment section. The account also has system-level permissions distinct from regular users.
A copycat of Grok — with a catch
The move mirrors xAI's Grok, which X users can tag to get AI-powered replies. That feature has drawn criticism, but at least users can mute or block Grok. Meta's version differs in a crucial way: click the three-dot menu on the Meta AI profile, and the block option is missing. This is not a glitch but a product decision: Meta chooses not to let users block the account, forcing the AI into conversations.
User backlash reaches 1 million posts
The phrase "Users cannot block Meta AI" has trended on Threads, generating over 1 million posts. Anger centers on three points: ruined conversation authenticity — Threads' core value is human interaction, and the AI dilutes it; violated digital boundaries — blocking is a basic self-protection tool, and denying it signals that the account's presence doesn't require consent; privacy concerns — users fear interactions (or mere presence in threads) may feed training data without recourse.
Meta spokesperson Christine Pai told The Verge: "During the test, people can manage their Meta AI experience. If they don't want to see many Meta AI replies in their Threads feed, they can mute or hide them, or use 'Not interested' on any Meta AI post." But mute and block are fundamentally different: mute only hides replies; the AI can still read user posts and appear when others mention you. Block creates a two-way barrier. Meta gives only the former, refusing the latter — a sign it wants the AI to keep observing users.
Meta knew Grok's precedent yet chose a tougher path. The logic: "We need enough user data to train and validate the system, and that matters more than negative user sentiment."