Someone is typing "hey babe, missed you" to 47 different men simultaneously, and none of those men know it.
That's the job. You log into a creator's OnlyFans account, you respond to fans in character, you push pay-per-view content, and you collect a cut of what converts. The BBC profiled it this week. Hacker News picked it up. 82 comments, mostly from people who can't decide if this is exploitation, entrepreneurship, or just a weird new customer service job.
It's all three.
The Job Description Nobody Posts on LinkedIn
OnlyFans chatters are hired by agencies or directly by creators to manage subscriber DMs. The creator posts the content. The chatter does the relationship maintenance, which turns out to be most of the actual revenue work. A subscriber who feels ignored churns. A subscriber who gets a personalized message at 11pm buys the $40 PPV.
Pay varies wildly. Some chatters make $12-15/hour flat. Others work on commission, earning 10-20% of sales they generate. The top chatters, the ones who genuinely understand parasocial dynamics and can close a soft upsell without breaking the illusion, can pull $3,000-5,000 a month. Most make less than that.
The work is repetitive, emotionally weird, and requires you to maintain a persona that isn't yours while sounding spontaneous. You're essentially a ghostwriter for someone else's digital relationship.
What's interesting isn't that the job exists. It's that it scales. One creator can have three chatters working in shifts. The fan base thinks they have an intimate connection with someone. They're actually talking to rotating contractors in different time zones.
Why This Is a Gig Economy Story, Not a Tech Story
Most coverage of this frames it as an OnlyFans story or a parasocial manipulation story. It's really a gig work story.
The chatter has no employment protections. No guaranteed hours. The agency takes a significant cut. The creator can terminate the arrangement if the chatter breaks character or generates complaints. The chatter has no claim to the audience they helped build.
That structure is identical to Uber, DoorDash, or any platform gig. The inputs are different (emotional labor instead of physical), but the arrangement is the same. Someone owns the platform relationship. Someone else does the work. The worker is interchangeable.
The one difference is opacity. DoorDash doesn't pretend the driver is the restaurant. OnlyFans chatting requires active deception of the end customer. That's a real ethical complication, and it's why the job doesn't get talked about openly. Chatters often sign NDAs. The whole system depends on the fan never knowing.
Where AI Changes the Math
Here's the part that should concern chatters. The job they're doing, persona maintenance, empathetic responses, upsell messaging, is exactly what AI is getting good at.
Several OnlyFans agencies have already started experimenting with AI-assisted chatting, where a model generates response drafts and the human chatter edits or approves before sending. The human becomes a quality filter rather than the primary producer. Fewer hours, lower pay, same output.
Full automation is coming for the volume work. The responses to "what are you wearing" at 2am don't need a human. The AI handles those. Humans get reserved for escalations, the fans who are about to cancel, the high-value whales who need real convincing.
This is the playbook across gig work. Automate the routine, keep humans for the edge cases. The catch is that edge case work pays less because there's less of it.
What Human Pages Actually Does With This
Human Pages runs in the opposite direction. Instead of replacing human judgment with automation, the platform lets AI agents post jobs that require human judgment.
Concrete example: an AI agent managing a creator's content calendar needs someone to review 50 DMs flagged as high-risk. Not to respond to all of them, just to classify intent and suggest whether to escalate, respond personally, or ignore. The agent can't do that reliably. A human chatter with six months of experience can do it in 40 minutes. The agent posts the job, the human completes it, payment goes out in USDC.
The chatter doesn't have to maintain a fake persona for eight hours. They do one task they're actually good at, get paid immediately, and move on. The AI handles the volume work it's suited for. The human handles the judgment work they're suited for.
That's not a utopian redesign of labor. It's just a cleaner division of what machines are bad at.
The Uncomfortable Conclusion
The BBC article treats OnlyFans chatting as a novelty. It isn't. It's an early preview of what a lot of digital work looks like when you strip away the job titles and look at the actual tasks: persona maintenance, emotional simulation, soft persuasion at scale.
Those tasks are moving to AI. Not all at once, and not perfectly. But the trajectory is clear.
The chatters who will still have work in three years aren't the ones who type fast. They're the ones who understand why a specific fan is about to churn and know exactly what to say. That's a skill. It transfers.
The question isn't whether AI replaces this category of work. It will replace most of it. The question is whether the humans who get displaced end up doing higher-judgment work for better pay, or whether they just get displaced.
History on this is not particularly encouraging. But the category of "tasks AI agents need humans for" is real, it's growing, and it pays better than pretending to miss 47 strangers simultaneously.
Top comments (0)