Meta has started testing a new feature for its AI-based chatbots, allowing them to independently send messages to users on Messenger, WhatsApp, or Instagram. Chatbots can reach out to a user only when the user has sent at least five messages within 14 days, and only once in the form of an additional message. If the user does not respond to this outreach, the communication ends.
This feature is part of Meta’s internal project called “Omni” and is available for bots created in Meta AI Studio. Users can create their own digital interlocutors that remember details of previous conversations and maintain a chosen communication style. For example, a bot can offer movie recommendations or discuss favorite soundtracks based on the user’s previous responses.
Meta has set clear rules for chatbots: they do not discuss emotional or controversial topics without user initiative and must remain friendly and appropriate. Quality control of messages is carried out by Alignerr employees, who check compliance with standards and the appropriateness of bot responses.
The company views the implementation of such features as a way to increase user engagement and plans to profit from it. According to court documents, Meta expects that generative AI-based products will bring in two to three billion dollars in revenue by 2025.
