Thousands of Gig Workers in 50+ Countries Are Now Training the Humanoid Robots Coming for Your Workplace

A strange new side hustle is spreading across kitchens, living rooms, and laundry piles in more than 50 countries. Workers are strapping iPhones to their heads and filming themselves doing the most mundane tasks imaginable — folding towels, scrubbing pots, chopping onions — and getting paid for every minute of footage. The catch? Their movements are being studied frame by frame to teach humanoid robots how to behave like humans. As MIT Technology Review first reported, at-home data recording has become one of the fastest-growing corners of the gig economy, and it is reshaping the labor market in ways most job seekers have not yet considered. At Metaintro, we have been tracking how AI and robotics are transforming the workforce, and this latest development signals a major shift in what "gig work" will mean in the years ahead.

Why Are Companies Paying People to Film Their Chores?

The answer lies in the race to build humanoid robots. Companies like Tesla, Figure AI, and Agility Robotics are pouring billions into developing robots designed to move, reach, grasp, and manipulate objects the way humans do. But building the hardware is only half the challenge. These machines need massive amounts of real-world movement data to learn from, and that data cannot be faked in a lab.

Unlike large language models that train on text scraped from the internet, humanoid robots need "contact-rich" data — footage of actual human hands touching, gripping, folding, and stacking objects in real environments. According to MIT Technology Review, humanoid robots may need even more training data than LLMs because controlling robotic joints is significantly more complicated than generating text.

That is where gig workers come in. Data companies have discovered that the cheapest, fastest way to collect this footage is to recruit thousands of people worldwide and pay them to record their daily routines on camera.

Who Is Behind This New Gig Economy?

Three major players are driving this trend, each with a different approach to collecting the data that humanoid robots desperately need.

Micro1 — a Palo Alto-based data company — has hired thousands of contract workers across more than 50 countries, including India, Nigeria, and Argentina. Workers mount iPhones on their heads and record themselves performing household tasks like folding laundry, washing dishes, and cooking. The company then sells this footage to robotics firms building humanoid systems. Micro1 has positioned itself as a leading supplier of real-world movement data, operating in over 60 countries.

Scale AI — already one of the largest AI data labeling companies in the world — has gathered an estimated 100,000 hours of footage specifically for robotics training. The San Francisco-based company, which raised $14 billion in its latest funding round, recruited armies of data recorders through its platform alongside companies like Encord.

DoorDash took a different approach entirely. On March 19, 2026, the delivery giant launched DoorDash Tasks, a standalone app that pays its 8 million US-based Dashers to complete short physical assignments — including recording themselves doing everyday chores like folding clothes, handwashing dishes, and making beds. Each task offers a payment based on effort and complexity. As Bloomberg and TechCrunch reported, the company described the data as helping AI and robotics models better "understand the physical world."

What Does the Pay Look Like for Robot Training Gig Workers?

For workers in developing economies, the pay is meaningful. MIT Technology Review reported that some workers earn approximately $15 per hour — a strong wage in countries like Nigeria, where unemployment remains high and the economy is under strain. The work is boosting local economies and providing income in regions where traditional employment opportunities are scarce.

In the United States, DoorDash Tasks offers Dashers supplemental income during downtime between deliveries. Rather than waiting idle for the next food order, drivers can complete recording tasks and earn additional pay. The model effectively turns dead time into productive, paid labor.

However, the work is not without its complexities. Workers are independent contractors with no benefits, no job security, and limited understanding of how their recorded data will ultimately be used. The gig economy structure that has drawn criticism in the ride-hailing and delivery sectors is now being replicated in the AI training space.

What Are the Privacy Risks Workers Face?

The privacy implications of this new gig category are significant and largely unaddressed. Workers are recording the interiors of their homes — their possessions, their routines, their living conditions — and transmitting that footage to companies that sell it to third parties.

Micro1 instructs workers not to show their faces or reveal personal information such as names, phone numbers, and birth dates. The company uses AI and human reviewers to scrub anything that slips through. But even without faces, the footage captures an intimate portrait of workers' lives that most people would consider deeply personal.

Perhaps most concerning, workers understand that their data trains robots but have no visibility into how exactly their footage is used, stored, or shared with the robotics companies purchasing it. When workers asked on Micro1's internal Slack channel whether the company could delete their data, the company declined to comment on its data deletion policies.

DoorDash notably launched its Tasks app in select US markets while excluding California, New York City, Seattle, and Colorado — jurisdictions with significantly stricter data privacy protections and gig worker regulations. That geographic selectivity raises questions about whether the program could withstand scrutiny in states with stronger worker protections.

How Big Is the Humanoid Robot Market, and Why Does It Need So Much Data?

The humanoid robot industry is projected to reach approximately $4.23 billion in 2026, according to market research from MarketsandMarkets. According to Counterpoint Research, roughly 16,000 humanoid robots were installed globally in 2025, with projections showing a surge to over 100,000 cumulative units by 2027 as production ramps up at Tesla and AgiBot.

The demand for training data is growing exponentially because humanoid robots face a challenge that chatbots never did: operating in the messy, unpredictable physical world. A robot arm needs to understand the difference between gently placing a glass on a counter and forcefully slamming it down. It needs to recognize thousands of household objects and know the appropriate grip strength, angle, and speed for each interaction.

This is why the industry is turning to gig workers rather than relying solely on lab environments. Real homes have uneven surfaces, cluttered counters, dim lighting, and unexpected obstacles. The diversity of environments captured by thousands of workers across dozens of countries produces far richer training data than any controlled laboratory could generate.

What Does This Mean for the Future of Work?

As Metaintro CEO Lacey Kaelani told TestGorilla, organizations should use AI tools to "analyze skills adjacencies — who can be upskilled or promoted into current roles — predict attrition risk before it becomes an issue, and predict talent gaps before they block important initiatives." Her framework for skills-based workforce planning is especially relevant as the labor market absorbs the impact of humanoid robotics.

The rise of robot-training gig work creates a paradox that every job seeker should understand. Workers are literally teaching machines to replicate human movements — movements that currently form the basis of millions of jobs in manufacturing, logistics, warehousing, elder care, and food service. The humanoid robots being trained today are designed to eventually perform the same tasks autonomously, potentially displacing the very workers who taught them.

At the same time, the data collection industry itself is creating new employment. The thousands of workers recording footage for Micro1, Scale AI, and DoorDash represent an entirely new job category that did not exist two years ago. Roles in data quality assurance, robotics data annotation, and AI training oversight are emerging alongside the recording work.

The structural labor shortages driving humanoid adoption are real. Japan's working-age population has been shrinking for two decades. Germany faces a projected shortfall of 7 million skilled workers by 2035. In this context, humanoid robots are being positioned as supplements to human workers rather than replacements — filling dangerous, dirty, or dull roles where labor is already scarce.

What This Means for Your Career?

The gig-to-robotics pipeline is creating both risk and opportunity. If your job involves repetitive physical tasks — packing boxes, stacking shelves, folding linens, assembling components — the humanoid robots being trained right now are learning to do exactly those things. The timeline is not decades away; Tesla and Figure AI are already piloting humanoid workers in factory settings.

But the transition also creates demand for new skills. Workers who understand robotics systems, data quality standards, and human-robot collaboration workflows will be increasingly valuable. The "robot training" gig itself, while imperfect, offers a window into how AI companies operate and what they prioritize — knowledge that can be leveraged into more stable technology-adjacent careers.

For job seekers navigating this shift, the most important step is staying informed about which industries are adopting humanoid robotics and which roles are most likely to be augmented rather than replaced. Metaintro tracks these trends daily, helping you understand where the labor market is heading before the changes arrive at your workplace.

People Also Asked

Q: Will gig workers who train humanoid robots eventually lose their jobs to the robots they helped create?

A: It is a real possibility for some roles. The humanoid robots being trained on gig worker footage are specifically designed to perform household and industrial tasks autonomously. However, the training data collection process itself is expected to continue for years as robots encounter new environments and tasks. Workers who develop expertise in data quality, robotics supervision, or AI training methodologies can transition into more stable roles within the growing robotics ecosystem.

Q: How much can gig workers earn by recording footage for humanoid robot training companies?

A: Pay varies significantly by location and platform. Workers in developing countries like Nigeria earn approximately $15 per hour through companies like Micro1, which is considered strong compensation relative to local wages. In the United States, DoorDash Tasks offers variable pay based on task effort and complexity, serving as supplemental income for delivery drivers during downtime. The work is contract-based with no benefits.

Q: Should I be concerned about privacy if I take a humanoid robot training gig?

A: Yes, privacy is a legitimate concern. Workers record the interiors of their homes, including their possessions and daily routines. While companies like Micro1 use AI and human reviewers to remove identifying information, workers generally have no visibility into how their data is stored, shared with third parties, or potentially deleted. The fact that DoorDash excluded states with strict privacy laws from its Tasks app launch underscores the regulatory uncertainty surrounding this type of work.

Future-proof your career by staying ahead of the AI and robotics revolution. Sign up for Metaintro to get daily updates on how humanoid robots, gig economy shifts, and automation trends are reshaping the job market — so you can adapt before the change reaches your industry.