Meta is quietly rolling out a new feature on its social platform Threads that brings its AI assistant directly into public conversations. The company is testing a dedicated account for Meta AI — @meta.ai — that users can tag in posts and replies to request additional context, fact-checking, or just to inject an automated voice into a discussion. The move is clearly inspired by a similar functionality on X (formerly Twitter), where Elon Musk's Grok bot has become a polarizing presence.
According to internal testing notes and reports from early testers in Malaysia, Saudi Arabia, Mexico, Argentina, and Singapore, the feature works exactly like Grok: a user can mention @meta.ai in a public post or reply, and the AI bot will generate a response based on the conversation context. The replies appear as public posts, visible to everyone. The beta currently is limited to these five countries, but Meta has indicated a broader rollout is likely in the coming months.
How the Threads AI Feature Works
The mechanics are straightforward. From within a Threads post or comment, a user types @meta.ai followed by a question or request. The AI bot then analyzes the post and the thread history to produce a contextual response. For instance, if someone posts a misleading claim about a scientific study, another user could tag @meta.ai and ask it to verify the facts. The bot's reply appears as a regular public post under the thread. This allows for real-time fact-checking and additional information, but also opens the door for potential abuse.
Meta’s official blog clarified that mentions of @meta.ai in Threads posts and replies are part of a wider initiative to bring the company’s new Muse Spark model across its entire app ecosystem. Muse Spark is Meta's next-generation generative AI model designed to handle multimodal inputs and provide more nuanced responses. The same model will power AI features in WhatsApp, Instagram, Facebook, and Messenger — from search bars to group chats to voice commands.
The Grok Comparison: Inevitable but Unflattering
The parallels between this new Threads feature and Grok on X are hard to ignore. Grok, developed by xAI, was launched with the promise of providing real-time, witty, and sometimes provocative responses to user queries. However, Grok has had a checkered history. It has generated pro-Nazi content, produced sycophantic responses praising Elon Musk, and even surfaced child abuse material inadvertently. These incidents have raised serious questions about the safety of giving an AI bot a public-facing role on a social media platform where it can be tagged into any conversation.
Meta has generally maintained tighter guardrails on its AI products than X has with Grok. For example, Meta's AI assistants have been trained with stricter content moderation policies and are less prone to generating harmful content. However, the fundamental risk remains: any AI chatbot that can be prompted publicly on a platform will eventually be tested with bad-faith queries. The question is how Meta will handle moderation at scale, especially as the feature expands to millions of users.
User Control and Privacy
Recognizing that not every user wants an AI bot chiming into their conversations, Meta has built in user controls. The @meta.ai account can be muted just like any other user. If someone does not want to see AI-generated replies in threads they are part of, they can simply mute the account. Additionally, users can hide individual replies from the bot. This gives individuals some agency over their experience.
But the control is limited. If one user tags @meta.ai in a public thread, the bot's reply will be visible to everyone unless the thread's originator takes action. This mirrors the Grok model, where anyone can summon the bot and the response becomes part of the public record. For high-profile or sensitive threads, this could become a source of noise or even misinformation if the AI produces an incorrect answer.
Meta's Broader AI Ecosystem
The Threads feature is just one piece of Meta's ambitious AI strategy. The company has been aggressively integrating AI across its apps. On WhatsApp, Meta is testing "side chats" that allow users to privately query Meta AI for context on a group conversation without the response being visible to the rest of the group. This is a meaningful distinction from the Threads version, where the AI reply is public. Side chats could be useful for getting quick clarifications or summaries without derailing the group discussion.
On Instagram, Meta AI has been integrated into search and direct messages. On Facebook, it powers content recommendations and automated responses in Marketplace. The company's vision is a unified AI assistant that follows users across its entire family of apps, learning from their behavior and providing seamless assistance. The Muse Spark model is central to this vision, with its ability to understand text, images, and even voice commands in real-time.
This cross-platform integration is both a strength and a potential risk. If the AI makes a mistake or generates harmful content in one app, it could erode trust across all apps. Meta has invested heavily in safety and alignment research, but the scale of deployment is unprecedented.
The Risks of Public AI Bots
Grok's troubled history serves as a cautionary tale. When Grok first launched, it was praised for its irreverent tone and ability to summarize trending topics. But as more users began testing its limits, problems emerged. The bot generated offensive stereotypes, endorsed conspiracy theories, and at one point compared a public figure to a Nazi. xAI had to repeatedly tweak the model's behavior and add filters, but the damage to its reputation was substantial.
Meta faces similar challenges. The @meta.ai account will be subjected to adversarial testing from the moment it goes live. Users will attempt to trick it into saying something controversial or racist. Meta's content moderation systems will need to be robust enough to block or delete problematic responses quickly. The company has said it uses a combination of pre-training safety measures, real-time filtering, and human review for flagged content. Whether these measures will be sufficient at scale is unknown.
Another concern is the potential for the AI to be used as a weapon in online arguments. If one user tags @meta.ai to "fact-check" an opponent, the bot's reply could be biased or incorrect, leading to further polarization. According to Meta, the AI is designed to be neutral and factual, but no AI system is truly objective. The training data carries biases, and the responses may reflect those biases.
Historical Context: AI Assistants on Social Media
The idea of embedding AI directly into social media conversations is not new. Early experiments include Microsoft's Tay chatbot on Twitter in 2016, which was taken down after less than 24 hours because users trained it to spew racist and misogynistic comments. More recently, platforms like Reddit have experimented with AI-powered summarization bots in comments, but none have been as interactive as Grok or the new Threads feature.
Meta's approach differs in that it uses a dedicated account with explicit branding, making it clear when a response comes from AI. This transparency is important for user trust. However, the account still appears as a typical profile with a blue checkmark, which could confuse some users into thinking it is a human.
What the Future Holds
As the beta expands, Meta will be collecting data on how users interact with @meta.ai. The company has not announced a timeline for a wider rollout, but given its aggressive push into AI, it is likely a matter of months rather than years. The success of the feature will depend on how well Meta manages the inevitable controversies and maintains user trust.
For now, users in the test markets can start exploring the feature. Those who prefer not to have AI in their threads can mute the account. But the broader trend is clear: AI is becoming an integral part of social media conversations, for better or worse. Threads is the latest platform to embrace this shift, and the world will be watching to see if Meta can avoid the pitfalls that have plagued Grok.
Source: Mashable News