Meta's Threads AI Account Can't Be Blocked by Users
Meta is testing a Threads feature that lets users tag a Meta AI account to get answers inline, but unlike every other account on the platform, users have no ability to block it. The asymmetry is the point.
Original sourceMeta has begun testing a dedicated Meta AI account on Threads, allowing users to @-mention it directly in posts and replies to get AI-generated answers. The feature mirrors how people tag other users for input, positioning Meta AI as a social participant rather than a settings-menu utility. On the surface, it looks like a natural extension of how people already use Q&A dynamics on social platforms.
The critical detail is what Meta quietly excluded: the ability to block the account. Every other account on Threads — brand, bot, or person — can be blocked. The Meta AI account cannot. Meta has not offered a detailed explanation, but the practical effect is that once this feature is fully deployed, any Threads user can summon AI responses into any conversation, and no other participant in that thread can remove it from their experience.
This isn't a technical limitation. Blocking is a standard content-control primitive that Meta already ships. Choosing not to extend it to their own AI account is a deliberate product decision — one that prioritizes AI distribution over user control. It also raises a structural question about consent: if someone @-mentions Meta AI in a thread you're part of, you will see that response whether you want AI in your feed or not.
The move is consistent with Meta's broader strategy of embedding AI across all of its surfaces — Facebook, Instagram, WhatsApp, and now Threads — with engagement metrics as the primary success signal. Whether regulators, particularly in the EU under the Digital Services Act, will treat an unblockable AI account as a design pattern worth scrutinizing remains an open question.
Panel Takes
The Skeptic
Reality Check
“An unblockable account on a platform where blocking is a core safety and preference tool isn't a feature gap — it's a policy choice dressed up as an omission. The specific scenario where this breaks: a user who has carefully curated their Threads feed to avoid AI-generated content gets it injected anyway because someone else in their thread tagged Meta AI. What kills this in 12 months isn't user backlash — it's a DSA enforcement action or an iOS App Store complaint forcing Meta to retrofit a toggle they could have shipped on day one.”
The PM
Product Strategy
“The job-to-be-done here is 'get an answer without leaving Threads,' and that's a real job — tagging a knowledgeable account for a quick answer is already native user behavior. But a product that solves one user's job by removing a control from every other user in the thread doesn't have a coherent JTBD — it has a distribution mandate. The gap between what's shipped and what's needed is a simple preference toggle: let users opt out of seeing Meta AI responses in their feed. That this wasn't included on day one tells you the product decision was made by growth, not product.”
The Futurist
Big Picture
“The thesis Meta is betting on: that social AI works best when it's ambient and impossible to fully opt out of, because optionality reduces network density and therefore answer quality. The second-order effect nobody is talking about is norm-setting — if the largest social platforms establish 'AI accounts are unblockable' as a design pattern, every platform that follows will treat it as precedent, not a choice. Meta is late to conversational AI but early to the specific power move of making AI a structural layer of social graphs rather than a feature, and that's the more consequential bet.”
The Designer
UX & Aesthetics
“Block is a control users reach for when they've lost trust or want to define the boundary of their experience — it's one of the most semantically loaded interactions on any social platform. Designing an account that looks like every other @-handle but silently removes that affordance is a failure of honest interaction design: the control appears to exist but doesn't, which is categorically worse than never offering it. The specific interaction failure here isn't the absence of a block button — it's that nothing in the UI signals to users that this account operates under different rules than every other one they interact with.”