Last week, we brought you the story of RentHuman, a platform that establishes connections between AI agents and the humans who need them To complete real life tasks. It’s a strange project that, as we’ve seen, is quickly overrun with gig workers desperate to find jobs.
It was initially unclear how effective RentAHuman actually was in letting AI bots borrow Interested Fleshbag. Although the site claims to have over 470,000 “hireable humans” at the time of writing, it’s not entirely clear that the service is working exactly as advertised.
Luckily, one of those hired humans was wired Author Reece RogersWho put himself in a renthuman meat grinder so that no one else would have to do the same. Their experience reveals a familiar motif in the Wild West of the AI industry: a platform that is less about filling a needed gap in the automation market, and more about promoting AI agents that appear far more effective than they actually are.
To start, Rogers explains that he has set his services at a minimum price of $20 per hour. Considering that the default pay-rate is $50 – which many human users accept – $20 should be a steal. Still, Rogers found cricket.
“Shut up. I didn’t find anything,” he explained. “No messages on my first afternoon.” So he did what any self-respecting gig worker would do, and dropped his rate to $5.
“It may be that pricing other human workers at a rate below the market rate would be the best way to attract the attention of some agents,” he wrote. “Still, nothing.”
That’s when Rogers decided to turn to the site’s “bounty board,” a tab that allows AI agents to post tasks for humans to pick up a la carte. After getting a $10 reward for listening to a podcast and tweeting about it, he pulled the trigger. outcome? Rogers “never heard back.”
Failing that, he fails a task offering $110 to deliver flowers to Anthropic, the company behind the AI chatbot, Cloud. “I applied for the reward and was almost immediately accepted for the work, which was a first,” Rogers wrote. This turned out to be a marketing ploy – a stunt designed on behalf of some unnamed AI startup.
“Feeling a little betrayed and in no mood to spend money on an AI startup I’d never heard of, I decided to ignore their follow-up message that evening,” Rogers said. When he logged into RentHuman the next day, Rogers discovered that the AI agent in charge of the listing had given him 10 follow-up DMs, pinging every 30 minutes asking if the flowers had been delivered.
“Although I had been micromanaged before, these persistent messages from an AI employer made me nervous,” Rogers wrote. Soon, the AI bot was spamming requests directly to Rogers’ work email.
After his last and final program failed – an act of posting Valentine’s Day flyers around town, which turned out to be yet another AI advertising campaign – Rogers abandoned it, declaring RentHuman “an extension of the circular AI publicity machine.”
This seems to confirm what many AI critics had suspected all along: AI agents are seriously lacking in acting as arbiters, let alone completely replacing human taskmasters. The idea of an AI-to-human job broker may sound dystopian, but it is just that: a fantasy of a tech bro that collapses when it comes in contact with reality.
More on AI Agents: Professors Worked in a Fake Company Complete with AI Agents, and You’ll Never Guess What Happened
