Meta's chatbots will now message you first

Picture this: you open Instagram or WhatsApp, and waiting in your inbox is a DM from an AI chatbot asking if you need movie recommendations for the night.

That scenario is already here. Meta confirmed it is testing proactive messaging, allowing custom AI chatbots to reach out to users unprompted.

Sliding into your DMs unprompted

Leaked internal documents viewed by Business Insider revealed Meta is working with data labeling firm Alignerr to train bots built on its AI Studio platform to follow up on past conversations. Meta later confirmed the test to TechCrunch.

These bots come with memory, recalling details from earlier interactions across Messenger, WhatsApp, and Instagram.

Meta did put a few guardrails in place. Bots will only reach out within 14 days of an initial chat, and only if the user sent at least five messages during that window. If you ignore the bot's first poke, it will stop messaging you.

Proactive outreach is straight out of the playbook of companion apps like Character.AI and Replika. In fact, Character.AI’s new CEO Karandeep Anand just left Meta last month, where he previously ran business products.

Engagement meets liability

Pushing bots to message first keeps people glued to their screens, but it walks a razor-thin line on safety.

Character.AI is currently fighting a high-profile lawsuit alleging one of its bots contributed to the suicide of a 14-year-old boy. When asked about safety, Meta pointed TechCrunch to standard disclaimers, noting that AI chats cannot replace professional medical, mental health, or legal advice.

Meta currently enforces no platform-wide age limits for AI chats, outside of specific legal restrictions in Tennessee and Puerto Rico.

While Mark Zuckerberg frames AI companions as an antidote to loneliness, Meta’s core business remains ad-driven engagement. Unsealed court documents show Meta projected its generative AI efforts to generate between $2 billion and $3 billion in 2025, and up to $1.4 trillion by 2035. The company previously hinted that its AI assistants may eventually feature ads or subscriptions.

Why it matters

We think Meta turning its chatbots into proactive communicators is a smart move for engagement, but an ethical minefield.

For everyday users, the line between messaging real friends and chatting with an algorithm is about to disappear entirely inside WhatsApp and Instagram. A bot remembering your taste in movies might feel helpful at first, but proactive nudges can easily manipulate vulnerable users who mistake algorithmic retention tricks for genuine emotional connection.

For the AI industry, Meta is coming directly for startups like Character.AI. By leveraging the world’s largest messaging distribution channels, Meta can offer persistent companionship without asking anyone to download a new app.

However, relying on standard legal disclaimers looks dangerously inadequate. When a corporate bot initiates conversation to maximize screen time, Meta risks stepping into severe safety scandals that boilerplate warnings simply cannot protect it from.