Menu
AI Data-Labeling Work
Each point below names the source it comes from and what that source actually says.
The basic gate is clearable — Outlier's terms ask for legal working age — but the real entry depends on more: Outlier's own FAQ lists at least an associate degree for its contributor path, and says pay rates vary by project, expertise, and location rather than a posted hourly.
This isn't a vague worry: reporting describes workers deactivated with thousands of dollars of finished work left unpaid, litigation against the platform's parent company alongside a large layoff, and an annotation model built on unpaid intro courses and qualification exams before paid tasks even begin. The pattern across these accounts is interruption and limited recourse.
Official copy says rates vary by project; worker reports show the realized money interrupted by unpaid training, tasks that dry up, deactivation, and payment disputes. There's no honest hourly to quote — which is why this reads as irregular side cash, not a wage you can plan around.
It's sold as a way into AI, but nothing in the evidence shows labeling work converting into AI or tech jobs — the tasks are anonymous and interchangeable, and employers don't read them as engineering experience. Treat the "into AI" framing as marketing, not a documented path; the useful residue is a feel for how the systems behave, not a credential.