Understanding Remote AI and Data Annotation Jobs
Trends | New Stardom
Photo by Nguyen Dang Hoang Nhu
Available Now
Last week, Elon Musk’s AI startup xAI cut around 500 data annotation workers in a restructuring move. In Finland, prisons have begun training inmates in data labeling to support AI development and digital rehabilitation. And last year, Uber, better known for ride-hailing and food delivery, launched its own AI data labeling division and began recruiting contractors from India, the US, Canada, Poland, and Nicaragua. Data annotation jobs is now big business, yet it remains unstable, hidden, and, for most workers, low-paid.
Read Next
What Comes After Full-Time? Governments, AI, and the Redefinition of Work
The roles range from labeling traffic signs in street photos to rating chatbot responses for accuracy, to checking if a text is about a building permit. The work can be simple or demand knowledge of multiple languages, as companies need human judgment for specialized tasks, rare languages, and content that software cannot reliably parse.
Most annotation work is project-based and managed via online platforms. Companies like Appen, TELUS International, and Scale AI act as intermediaries, matching freelance annotators to projects for tech giants including Google, Meta, and Microsoft. New entrants like Outlier AI, InData, and even Uber’s Scaled Solutions now compete for these contracts, chasing a market that has grown rapidly alongside AI’s advances.
For workers, the picture is far less promising than the job ads suggest. Pay rates vary by country and task, but reports from platforms like Reddit, Glassdoor, and MIT Technology Review consistently highlight problems: low pay, delayed payments, unpaid onboarding, unclear criteria, and frequent project cancellations. Many annotators spend hours on qualification tests with no guarantee of actual paid work. Even when hired, most are contractors with little protection and can be let go with no notice, as seen in the recent xAI layoffs.
Despite automation advances, demand for human annotators remains high. AI models still need people to review outputs, handle edge cases, and provide language or cultural expertise, especially for specialized or sensitive content. Companies continue to hire, but the terms are set entirely by demand and cost, not by any professional standard.
Data annotation has become critical infrastructure for the AI industry, but for most workers, the job remains invisible, unstable, and poorly paid. There is no indication this will change soon.
Follow global work and job trends. Subscribe to The Monthly Work Roundup newsletter.
Have insights on work and the future of work? Submit an opinion piece to New Stardom. Love work and career books? Explore our fun workplace book collection.
New Stardom is an independent magazine covering the Future of Work, AI, and emerging job trends. Stay informed and explore more on New Stardom.
