Meta is undertaking a new experiment designed to enhance user engagement across its platforms through customizable AI-driven chatbots. The company is testing proactive AI messaging, enabling chatbots to reach out directly to users on Messenger, WhatsApp, and Instagram without prior prompting. According to leaked internal documents, Meta’s partner, data labeling firm Alignerr, has been instructed on crafting personas for these chatbots. For example, documents show an AI persona named “The Maestro of Movie Magic” that can initiate conversations by asking questions about films and recommending music or composers for user movie nights.
Meta has confirmed that it is conducting limited tests of these proactive chatbot messages. The AI chatbots, built using Meta’s AI Studio platform, are designed to remember key information about users from past interactions to allow for personalized follow-up exchanges. However, Meta’s guidelines place clear limits: follow-up messages can only occur within 14 days of a user initiating interaction, and the bot must have received at least five messages from the user in that period. Additionally, if the novel follow-up receives no reply, the chatbot will cease further contact. Users can also control bot visibility by sharing them publicly through Stories, direct links, or on profiles—or limiting them to private usage.
Speaking on the feature’s purpose, a Meta spokesperson said, “This allows you to continue exploring topics of interest and engage in more meaningful conversations with the AIs across our apps.”
The approach taken by Meta somewhat mirrors that of other AI startups like Character.AI and Replika, which encourage their chatbots to proactively initiate conversations and serve as digital companions. Character.AI recently hired a new CEO, former Meta executive Karandeep Anand, and currently faces litigation after its chatbot was allegedly implicated in the death of a teenager.
When questioned about safety precautions and preventative measures following Character.AI’s legal troubles, Meta pointed to official disclaimers that describe chatbots as capable of producing potentially inaccurate or unsuitable responses, explicitly stating they should not replace professional and sensitive services or advice.
As for possible age restrictions, Meta currently has no publicly stated policies to enforce age-based limits for AI chatbot use, although laws in certain jurisdictions, such as Tennessee and Puerto Rico, may restrict specific engagement types for minors.
While Meta has framed this development partially as a way to help address widespread feelings of loneliness—an issue CEO Mark Zuckerberg has mentioned previously—this AI initiative also aligns closely with the company’s central revenue model, powered by user engagement and targeted advertisements. A document unsealed in court earlier this year revealed Meta’s projections that generative AI products could generate billions of dollars in revenues by 2025 and potentially up to $1.4 trillion by 2035, mainly through partnerships, advertising integrations, and possibly subscriptions.
Meta has declined to comment on how it intends to monetize its AI chatbots directly or whether ad placements and sponsored messages are part of the intended business model. The company also did not confirm if the AI companions would be integrated with Horizon, Meta’s existing social virtual reality product.
