Meta has begun testing a way for Threads users to send private messages directly to its AI chatbot. The feature lets people share posts or images with Meta AI inside the app’s direct messages without those exchanges appearing on any public feed. The company describes the conversations as fully private, though the Meta AI account itself cannot be blocked. Users can still mute the profile or hide individual replies through existing feed controls.
This move fits a broader pattern. Meta has been inserting its AI assistant into more corners of its products for months. On WhatsApp and Messenger, people already summon it in group chats with a simple tag. The company appears determined to make the chatbot a constant presence across its network of apps. That urgency has not always served it well. Last month, attackers used basic prompt injection and location spoofing to convince Meta’s AI support tool to change recovery emails on Instagram accounts. High-profile profiles, including the official Obama White House account and that of a senior Space Force officer, were among those compromised. The incident showed how quickly an always-available AI can become a liability when security lags behind rollout speed.
Threads itself has spent much of the year adding tools meant to give users more say over what they see and how they interact. Early on came a public “Dear Algo” option that let people request temporary changes to their recommendation feed. A quieter “Your Algo” version followed, allowing the same adjustments without broadcasting them. Other updates included indented replies for clearer conversations, private likes, automatic splitting of long text into threaded posts, a Communities Hub with live chat, and the ability to select specific phrases in a post and respond with custom emoji. These changes suggest the platform is still trying to define itself against more established competitors, balancing algorithmic discovery with finer personal controls.
Placing an AI chatbot inside private messages continues that experiment. It offers a convenient place to ask questions or generate ideas without cluttering the main feed, yet it also keeps the company’s AI system one step closer to every conversation. Whether the private channel improves usefulness or simply multiplies the surface area for the same kinds of prompt-based failures seen on Instagram remains an open question. For now the test is limited, and the inability to block the account means the AI presence stays available whether users want it or not.
The expansion of Meta AI into Threads direct messages therefore sits at the intersection of convenience and caution. It extends a pattern of rapid deployment that has already produced both new features and notable security lapses. Users who value private interaction with the chatbot may welcome the option. Those who prefer clearer boundaries around automated accounts may find the permanent presence harder to ignore.
