AI Training Tasks (Data Annotation)

What it typically pays

Public sources report $10 to $30 per hour for ai training tasks (data annotation). DataAnnotation.tech roughly $10-$14/hour for general tasks and $15-$25/hour for coding; Outlier about $15-$30/hour standard and $25-$45/hour for coding (2026 worker reports, Glassdoor). Paid only for timed task work; project availability is volatile. Prices vary by city and by how the work is packaged. The launch checklist includes a pricing step so you set yours on purpose, not by guessing.

What it costs to start

Expect about $0 before the first job. Estimate (no single source): laptop and internet you already have The basics:

• Laptop
• Stable internet
• PayPal
• A writing style guide
• Timer

How fast the first dollar comes

People commonly report a first paid job within 7 to 30 days of starting, when they follow the checklist below in order. The fastest route is almost always someone you already know.

Where the first customers are

• Apply directly: DataAnnotation.tech, Outlier (Scale), Alignerr, Mercor; applications only, no clients
• Specialty queues (coding, medical, legal, non-English languages) pay the most
• Keep 2-3 platforms active to smooth dry spells

Rules, permits and risks

• Independent contractor: pay only for approved task time; accounts can be paused without appeal
• Strict confidentiality and no-AI-assist rules; violations end the account
• Volatile: weeks with no projects are normal; treat it as supplemental only
• Income is taxable whether or not you receive a 1099; set aside roughly 25-30% (federal income tax plus 15.3% self-employment tax) and pay quarterly estimates if you expect to owe $1,000+ (IRS Self-Employed Tax Center).

Launch checklist

1. Apply to DataAnnotation, Outlier, and one more with a clean writing sample
2. Complete the qualification assessments carefully; they gate everything
3. Read each platform's guidelines; keep a notes file of rubric rules
4. Work 2-hour focused blocks; track paid vs unpaid time
5. Qualify for specialty queues (coding, your language, domain expertise)
6. Keep quality scores high; skip tasks you are unsure about
7. Withdraw earnings weekly
8. Log monthly hours and effective rate; drop platforms under $12/hour effective
9. Use the income to fund a hustle with a ceiling; this one is a floor, not a ladder
10. Reassess after 90 days: keep, cut, or switch platforms

Open as a checklist you can tick off

Templates

Replace the [brackets] with your details and send.

Honest caveats

• No ceiling growth; it is a remote hourly gig
• Dry spells with zero work
• Screening is competitive; many applicants never get projects
• Sitting and screen fatigue