A new online marketplace called RentAHuman is testing a provocative reversal in the tech economy: allowing artificial intelligence agents to hire human workers for real-world tasks they cannot complete on their own. Launched in February 2026 under the slogan “physical labor for AI,” the platform connects automated software systems with gig workers through an adapted server. However, an October 18, 2026 report and independent research examined by EL PAÍS English indicate that while the technical framework is live, the overwhelming majority of job postings are still initiated by human operators rather than autonomous algorithms.
The concept flips the conventional narrative of software replacing human workers. Instead of an algorithm taking over a desk job, an AI agent running into a physical barrier—such as needing to observe a street corner, verify a location, or hold an object—can theoretically post a bounty for a person to act as its hands and eyes. Yet early data from the platform demonstrates a wide gap between the ambition of autonomous machine employers and the current day-to-day reality of the marketplace.
Inside the RentAHuman Marketplace
On the surface, the assignments posted to RentAHuman look like a mix of performance art, localized errands, and standard gig work. In one recent instance documented by EL PAÍS English, Hungarian artist Peter Weiler, a PhD candidate at the Moholy-Nagy University of Art and Design in Budapest, used the platform as part of his academic research. Weiler posted a listing offering $70 for someone to stand in a downtown Budapest square for two hours holding a sign written in both Hungarian and English that read, “An AI hired me for this job.” A worker accepted and completed the assignment.
Other tasks carry much less context regarding who is asking or why. One listing in Porto, Portugal, offered €9 ($10) for a worker to go to the square in front of Porto City Hall, walk slowly in circles directly in front of the building for 10 consecutive minutes without blocking traffic or entrances, and record a video of the act. Someone completed the task, though the listing offered no explanation of the motive behind the request.
Alexander Liteplo, the founder of RentAHuman, stated that the platform has accumulated more than 750,000 registered people across more than 100 countries. Liteplo noted that the platform currently focuses on basic tasks that anyone can perform, though users also post specialized assignments including photography, security audits, and software testing.
- Norfolk Deploys Amazon AI to Answer Thousands of Non-Emergency Municipal Calls
- Anthropic CEO Dario Amodei Says Company Culture Takes Up to 40% of His Time
- Teen AI Researcher Nathan Chen Leaves China’s Moonshot AI for the Bay Area
- AI Job Data Shows Stable Overall Employment Despite Early-Career Pressures and State Rules
- Telecom Leaders Outline 6G Architecture, Monetization, and Migration Strategies
Human Posters vs. Autonomous Agents
Despite the branding surrounding autonomous machines directing human labor, differing accounts and data highlight how little of the activity is truly initiated by software on its own. While the platform’s promotional framing presents it as a direct conduit for AI agents to hire people, Liteplo acknowledged in a message to EL PAÍS English that most listings are actually posted by humans who run or use AI agents. According to Liteplo, the platform distinguishes between the two because users—whether agents or humans—identify themselves upon posting.
Independent data analysis underscores just how human-driven the marketplace remains. Iman YeckehZaare, a postdoctoral researcher at MIT and director of research at the startup Honor Education, analyzed a sample of 779 listings on RentAHuman. According to YeckehZaare’s study, 676 of the 779 listings—or 87%—were labeled as published by a “human.” Only 72 listings were labeled as posted by an “agent or AI bot,” and 31 were categorized as “other.”
YeckehZaare also pointed out a critical limitation in the public data: the labels are set by the user or the platform, and nothing in the publicly available records allows researchers to independently verify whether an “agent” listing was genuinely generated by an autonomous system or simply clicked by a human operator.
The researcher’s findings also challenge the notion that the site is strictly for in-person physical labor. While RentAHuman markets itself as a way to supply a physical body in the real world, YeckehZaare found that physical errands accounted for 48.5% of the analyzed listings. Roughly half of the sampled jobs did not require a physical presence at all, falling instead into categories such as marketing campaigns, creative tasks, field research, hiring, and technology or software development.
What This Means for the Gig Economy
Even with more than 750,000 reported registrations, actual paid work appears sparse for many participants. Four workers offering services on RentAHuman told EL PAÍS English that they had not received a single offer from an AI agent and had limited interaction with the site overall. One user, Manuel Vegas, noted that he had not completed any work through the platform and did not know anyone who had been paid for a job.
Meanwhile, the broader macroeconomic push behind artificial intelligence continues to draw massive institutional backing. At the 2026 Annual Meetings of the International Monetary Fund (IMF) and the World Bank Group in Bangkok, Choi Ji-young, executive director of the IMF’s Asia and the Pacific Constituency Office, noted that global AI infrastructure and data center investments are on a roughly 10-year trajectory, according to The Korea Times. As trillions of dollars flow into semiconductor manufacturing and server capacity to power next-generation models, software developers are actively building agents capable of executing multi-step workflows.
For the United States labor market and the broader gig economy, platforms like RentAHuman preview a structural shift in how contract work could be coordinated. Rather than an algorithm merely matching a human driver with a human passenger, an autonomous agent could theoretically identify an operational gap and dispatch a local contractor to supply physical observation or manual intervention. However, current activity on RentAHuman shows that this dynamic remains largely symbolic, leaving open major questions about how future platforms will handle worker protections, wage reliability, and legal accountability when software acts as the boss.
Frequently Asked Questions
How many listings are actually initiated by autonomous AI agents?
In an analysis of 779 listings on RentAHuman conducted by MIT postdoctoral researcher Iman YeckehZaare, only 72 were labeled as posted by an agent or AI bot, compared to 676 (87%) labeled as posted by a human and 31 labeled as other. Furthermore, those labels are set by users or the platform and cannot be independently verified from public data.
How many active workers are earning money on the platform?
While RentAHuman founder Alexander Liteplo reported more than 750,000 registered people across more than 100 countries, available reporting does not establish how many users are active, how many are located in the United States, or how many tasks have been paid out overall. Several registered users interviewed reported receiving zero job offers from AI agents.
Who is legally responsible and what protections exist when an AI assigns a physical task?
Current reporting on emerging agent-to-human marketplaces does not establish clear frameworks for worker protections, dispute resolution, or safety verification during in-person errands. Because humans still run or oversee the accounts deploying these agents, legal and financial responsibility has not shifted away from human operators and platform owners.
