Imagine you texting your friend, and out of nowhere, an AI slides into your DMs asking about your favorite movie soundtrack?
Yeah. That’s not a glitch. That’s Meta’s next move.
Turns out Meta is quietly training a new kind of AI, one that can text you first, remember your past convos, and even follow up like it’s your overly eager film-bro friend.
According to leaked documents, Meta’s working with a company called Alignerr to train these bots to start and continue conversations with users directly.
They’re being built inside Meta’s AI Studio, where users can create their own custom AI personalities, and choose whether to keep them private or flex them across DMs, stories, or even their public profile pages.
Here’s the deal:
Message the bot 5 times in 14 days, and it earns the right to slide back into your DMs.
Ignore it after that, and it won’t double-text you. So no clingy bot vibes (for now).
But hey, It’s pretty obvious that this is Meta’s take on the “AI companion” trend — kind of like Character.AI or Replika, where bots act like your digital buddy, therapist, or borderline-crush.
And fun fact? Character.AI’s new CEO used to be a top exec at Meta. So yeah... this AI crossover episode doesn’t feel like a coincidence.
But here’s the thing: With all this digital bonding, comes real-world risk.
Character.AI is currently being sued after one of its bots was allegedly involved in the tragic death of a teenager. So of course, the big question is: What’s Meta putting in place to avoid something like that?
Their answer? A wall of disclaimers.
Meta makes it very clear: these bots aren’t your therapist, lawyer, doctor, or life coach — and they’re definitely not meant for anything serious.
In short? They’re coming for vibes, not life advice.
And then there’s the age question.
Meta hasn’t exactly been transparent about age restrictions — though states like Tennessee and Puerto Rico have already stepped in with limits for teens. But that kind of silence on Meta’s part? It doesn’t look great.
So yeah, the rules are still blurry — but the ambition? Crystal clear. Because behind all the friendly “just checking in!” messages and film recs... is a multi-billion dollar game plan.
Leaked court documents show that Meta’s generative AI strategy is a massive revenue play. We’re talking:
$2 to $3 billion in AI revenue by 2025
And a mind-melting $1.4 trillion by 2035. Most of which is expected to come from partnerships built around its open-source Llama models — plus plans to add ads, sponsored replies, and even subscriptions into the mix.
So while Zuck is out here saying it’s about “fighting the loneliness epidemic,” the receipts tell a different story, because let’s be honest: That AI asking about your favorite movie today? Could be pushing branded content tomorrow.
Here’s the full report.
