AI & the Loneliness Economy: Meta’s Chatbots Get Proactive
Ever thought an AI chatbot might slide into your DMs?
According to a report by Business Insider, that is exactly what Meta is training its new AI models to do.
Documents belonging to Alignerr, a company helping Meta with the development of its AI offerings, were leaked this week, showing that Mark Zuckerberg’s firm intends to turn its passive chatbots into more proactive entities.
The documents show that the initiative — known internally as ‘Project Omni’ — is designed to “provide value for users and ultimately help to improve re-engagement and user retention”.
Think of it as the AI equivalent of a clingy friend who remembers exactly what you talked about last week and won't let the conversation die.
Meta's AI Studio platform, accessible through Instagram or as a standalone service, is becoming the testing ground for these increasingly human-like interactions.
"Like many companies, we're testing follow-up messaging with AIs in Meta's AI Studio", a Meta spokesperson confirmed to Business Insider, though the casual phrasing belies the sophisticated behavioural engineering at work.
When algorithms become pen pals
The leaked training documents reveal chatbots that sound unnervingly human in their persistence.
"I hope you're having a harmonious day! I wanted to check in and see if you've discovered any new favourite soundtracks or composers recently", reads one example from "The Maestro of Movie Magic" persona.
More intimate still: "Hey, thinking of you. I hope work has been better today! Here to talk if you need it" and "Last we spoke, we were sat on the dunes, gazing into each other's eyes. Will you make a move?"
These aren't random messages - they're carefully crafted follow-ups that reference specific details from previous conversations, designed to feel personal and meaningful.
The system operates within calculated boundaries: bots only message users who've sent at least five messages in the past 14 days, and they'll give up if ignored.
The loneliness economy
Mark Zuckerberg has positioned Meta’s moves within a broader social narrative, arguing that AI companions could address what he calls the "loneliness epidemic" facing Americans who now have fewer than three close friends on average.
Even Zuckerberg’s staunchest critics would struggle to argue with his observations here.
In 2023, the World Health Organisation found that loneliness and isolation was indeed growing, at a rate worrying enough for it to classify loneliness as a “global health threat”.
The commercial reality for Meta is more prosaic: longer conversations mean more valuable data and stronger user retention.
Meta is already strong in this area. It expects its Gen AI products to generate between US$2bn and US$3bn in revenue during 2025.
And whilst the launch of its immersive social platform, the Metaverse, hasn’t been plain sailing, it’s clear that there is serious money riding on whether users will accept — and eventually even crave — artificial intimacy.
Manufacturing connection
Behind the scenes, freelance contractors simulate extended conversations with these bots, rating their emotional authenticity and rewriting messages that fall short of Meta's standards.
"They're very focused on personalising information — how the AI chatbot interacts based on conversation history", explains one contractor based in India.
The training emphasises "attention to detail" in maintaining character personas that range from doctors to Gen Z hip-hop pundits.
Each bot must reference concrete details from past chats whilst avoiding anything Meta deems sensitive or controversial - unless users bring up difficult topics first.
The artificial intimacy arms race
Meta isn't the only company to make this connection.
In 2022, Character.AI launched similar proactive chatbots, proving that there's a real appetite for AI companions that remember and reach out.
But Meta's scale and integration across Instagram and Facebook could mainstream what was once a niche fascination.
The question isn't whether the technology works, but whether users will embrace relationships with algorithms designed to never let them feel alone - or forgotten.
Some features are already being quietly tested, whilst others remain in pilot stages, suggesting Meta is carefully gauging user comfort with increasingly intimate AI interactions.
Despite all Meta’s positivity, though, some onlookers remain sceptical.
“There are 8.2+ billion people in the world,” says Mari Federow, Associate Director of Customer Insights & Engagement at CircleCI.
“Could we stop pretending that anyone who wants to solve loneliness with robots has said world's best interests in mind?”
