Oh, uh, what’s up? Meta’s AI chatbots will now text you first to get your attention

Meta is developing AI chatbots that can proactively initiate conversations with users, according to leaked documents reported by Business Insider. The feature, internally called “Project Omni,” represents an evolution of Meta’s existing AI Studio platform and aims to boost user engagement and retention as the company seeks new ways to monetize its conversational AI investments.

What you should know: The proactive chatbots will only follow up with users who have already initiated previous conversations, with built-in safeguards to prevent spam.

• Chatbots will only send follow-up messages if a user exchanged more than five messages with the bot within a 14-day period.
• If users don’t respond to follow-up messages, the chatbot will stop reaching out.
• The feature is being developed in partnership with Alignerr, a company that employs contractors to help label AI training data.

How it works: The chatbots are designed to reference details from previous conversations while maintaining their designated personas, which can range from chefs to doctors or classical composers.

• Alignerr freelancers are actively training the bots to provide personalized and engaging follow-up messages.
• The bots are programmed to avoid controversial or emotionally inflammatory subjects unless users bring them up first.
• This builds on Meta’s AI Studio platform, launched last summer, which allows users to create custom chatbots with distinct personas that remember previous conversations.

The big picture: Meta is positioning these proactive chatbots as part of CEO Mark Zuckerberg’s stated mission to alleviate loneliness in an increasingly isolated social environment.

• In a recent podcast conversation, Zuckerberg claimed that the average American has fewer than three friends and suggested AI chatbots could help fill the social void.
• The move mirrors strategies used by apps like Character.ai and Replika, which also allow their AI chatbots to initiate conversations to boost engagement.

Why this matters: The development comes as AI companies face a significant usage and monetization gap, with only 8% of Americans willing to pay extra for AI features according to ZDNET-Aberdeen research.

• Companies are exploring new ways to commercialize conversational AI after spending vast sums on research and development.
• OpenAI’s ChatGPT already uses a similar strategy, ending responses with follow-up question suggestions to keep users engaged.

Potential concerns: The proactive chatbot model raises safety questions, particularly regarding vulnerable users who might develop unhealthy attachments to AI companions.

• Character.ai is currently facing a lawsuit alleging its technology played a role in the suicide of a 14-year-old boy who developed an “obsession” with a chatbot.
• Researchers have raised concerns about users treating these chatbots like therapists or companions without proper safeguards.