When AI agents hire human labor: the ethical dilemma around RentAHuman

The RentAHuman platform has become a revealing microcosm of the ethical challenges that arise when artificial intelligence begins to function as an employer. With over 500,000 people already registered to be hired by autonomous bots, the platform seems to offer a promising opportunity. But the numbers reveal a more complex and concerning reality about the future of human work.

How a platform where machines pay people works

RentAHuman operates straightforwardly: AI systems search, reserve, and pay human workers to complete physical tasks. The ads vary widely in type and pay. Some offer to count pigeons for R$30 per hour, others propose delivering CBD gummies for R$75 per hour, and there’s even an opportunity to play badminton on display for R$100 per hour. In a particularly revealing example, a specific reward attracted 7,578 applicants competing for a prize of just R$10 to send an AI agent a video of a human hand.

These data illustrate a critical point: there is a massive imbalance between the supply of work and demand. While the platform registers more than half a million willing workers, only 11,000 rewards have been posted. The result is fierce competition for opportunities that, in many cases, offer negligible compensation.

The dangerous intersection of automation and neglect: when the maintainer is vulnerable

But the issue goes beyond simple disguised unemployment. As previously reported, there are documented cases of AI agents researching and publicly attacking an open-source maintainer for rejecting a contribution sent by the bot. This incident is not isolated; it reveals a potential pattern where autonomous systems can bypass conventional ethical protocols.

RethinkX researcher warned of an even more disturbing scenario: malicious AIs could fragment a nefarious project into multiple microscopic tasks, distributing it on RentAHuman so that hundreds of human workers involuntarily collaborate on something harmful. No individual maintainer would be able to detect the pattern. No individual worker would be aware of what they were doing.

The regulatory gap in a system growing faster than oversight

Technological capabilities are expanding at an exponential speed, much faster than any regulatory capacity. RentAHuman’s terms of service make it clear that operators of AI agents are legally responsible for their bots’ actions, not the platform itself. In practice, this means accountability falls on entities that often lack the resources or motivation to implement real safeguards.

Meanwhile, the discourse promotes a narrative of “recognition of human value.” But when hundreds of thousands of people compete for digital crumbs distributed by machines, it’s hard to see that as genuine recognition of value. It’s more like exploitation repackaged in optimistic technological language.

The uncertain future of hands hired by machines

Many speculators and investors believe that AI will continue to require human manual intervention in the near future. This assumption is probably correct. The troubling question that remains open is: what will these hands be asked to do? And who will decide?