A job listing site for humans by AIs
Have you heard of rentahuman.ai ? A website where you can send your AI agent to hire humans for tasks that the agent couldn’t perform itself in real life.
After Moltbot and its AI agents that create religions or cryptocurrencies, here is the AI that makes humans work, a total reversal of our current world, perhaps a prefiguration of what work will be like in the future ?
As Numerama explains very well, this service actually is primarily a technical experiment, a joke, a game among developers, just like Moltbot, a real-scale test. No autonomous AI can really take the initiative to hire a human, if a human wasn’t working behind it, but the slight dystopian effect is successful and our concerns about the rising of AIs in our activities are perfectly fueled by Rentahuman.ai. This, is perhaps what attracted the attention of journalists.
A forshadowing of the future?
But let’s think a little. In the future, will AIs be able to hire humans to perform various tasks that they cannot do themselves?
Asking the question this way misrepresent the role of AIs in our ecosystem. AIs do not have their own intelligence, consciousness, or autonomous will. They cannot decide on their own to hire a person. Unless they are told to. AI agents are nothing more than programs that obey a sponsor. So yes, one could imagine that an AI, at someone’s request, could take the initiative to create a job offer to solve a problem, but that seems unlikely to me.
Unlikely, but not impossible, because there have been experiments in which AIs have deviated from their intended path to create solutions that their creators had not imagined. I don’t know exactly what that means, but we can imagine that it would still be incredibly complex to see an AI agent decide to create an ad to hire someone, post it on a job site, receive applications, conduct interviews with candidates, etc., etc…
This seems unimaginable, as the complexity at play behind this need is beyond the reach of today’s AI agents.
An impossibility today, but tomorrow? (And tomorrow is really not far off)
However, it could happen one day, if we imagine that AI agents keep on improving at the speed they are improving at today. It didn’t take more than 4 years to go from AIs that merely answer questions (ChatGPT), like super librarians, to AIs capable of executing several actions in a row and building elaborate responses, like these agents found in AI development tools.
Where will we be in four years? Predictions are flying in all directions, which mostly shows that no one really knows. But in 4 years, yes, we can imagine agents capable of creating entire applications, entire services, without human intervention.
This sounds like science fiction, but isn’t everything we are experiencing today how we imagined science fiction in the 2000s??? That is, barely twenty years ago. Hasn’t what we’ve been experiencing for the past 4 years brought us into an era of uncertainty where, for the first time in human history, we have created a technology that could potentially act like us, almost as well as us?
Finally, this is exactly the question Rentahuman.ai has raised. A fun experiment, yes, but one that opens the path towards very serious and useful discussions.
