A nonprofit worker earned about $31,000 in 18 months doing AI training tasks on Outlier after hours
Working 20 to 25 hours a week on top of a full-time job, he trained video chat models and recorded voice data for AI companies.
Subject: Ryan AdamsPaid AI training and data tasks as an independent contractor
Case Overview & Background
Ryan Adams, who works full-time at a health and environmental nonprofit in Virginia, found Outlier through a Google search while trying to learn about AI. Over about a year and a half he worked on about 30 projects, including training video chat models and voice recording, and earned about $31,000.
How AI Is Used
Step-by-Step Process
He applied to Outlier and got projects about three weeks later.
About 30 projects in 18 months, including video and voice work.
20 to 25 hours a week, up to 30 when bonus pay is offered.
Customer Acquisition & Distribution
Limitations, Risks & Failure Modes
What may not apply to you, and what could go wrong:
- There is no job security and projects can end suddenly
- Pay rates are set by the platform and can change
- The work can be monotonous and poorly run, other contractors in the source say
- His figure is self-reported
Key Takeaways
- AI training work can be done without technical skills
- Income depends on platform availability rather than your own customers
- It works best as a supplement, not a business you control
Sources
Business Insider: 5 AI training contractors share how they broke in and how much they made
Note: Ryan Adams section reports the about $31,000 total, hours and project types.
Category page banner: Reach readers researching one specific AI business model.
One clearly labelled banner on a category page such as Micro-SaaS or AI Agencies, shown next to the documented cases your product is relevant to.
Sponsored placements are always labelled and never change how cases are researched, labelled, or ranked.
Commercial disclosure: We currently have no affiliate or sponsorship relationship with any tool or person on this page. If that changes, affiliate links and sponsored content will be clearly labelled, and evidence labels will never be changed for commercial reasons.
