Expert network · Assessment-gated contractor

DataAnnotation

DataAnnotation is a curated, assessment-gated platform for educated generalists and licensed experts, with a strong writing-and-reasoning bent and one of the best payment-reliability reputations in the gig space. There is a single always-open application rather than job postings.

$15–$100+/hr

Platform-stated

Join DataAnnotation How to get accepted

Is DataAnnotation legit?

Researched October 2026 · 15 sources

Yes. DataAnnotation is the worker platform of Surge AI, a large, profitable San Francisco AI data company, and it pays approved work through PayPal. Your account or your work can vanish without a reason: reviewers describe locked accounts with earnings held, empty dashboards for months and support that rarely replies. The one-shot assessment and qualification tests are unpaid. It pays $15–$100+/hr (platform-stated).

Joined through its own application (it publishes no job postings)

Who is behind it

DataAnnotation is run by Surge AI (Surge Labs, Inc.), the San Francisco data-labeling firm Edwin Chen started in 2020. Forbes reported in September 2025 that Surge made $1.2 billion in revenue in 2024 without outside funding and hires annotators through DataAnnotation, a wholly owned affiliate whose site and job listings never name Surge. DataAnnotation's own FAQ says it connects over 100,000 contractors with AI companies.

Getting paid

DataAnnotation pays only through PayPal, to the email on your PayPal account. Its FAQ says deposits arrive within a few days after you request them, and its job pages list payouts as weekly. It charges no signup fees and says it will never ask you for money.

Worker ratings

Trustpilot 3.9/5 (1,974 reviews). Read October 2026.

Who can join

Its FAQ names no country list (it says it connects over 100,000 contractors globally), but the roles on its job board and worker reports point to six countries: the US, Canada, the UK, Ireland, Australia and New Zealand (read 18 Oct 2026).

Fake DataAnnotation recruiters

DataAnnotation warns that scammers use its name, says it never asks for money and that the only way to apply is its sign-up at app.dataannotation.tech. In a January 2025 BBB report, a fake 'Data Annotation Tech' task site on .cc lookalike domains, pitched by text and WhatsApp, took about $20,000 from a New Yorker. Never pay to get work.

What people praise

Most Trustpilot reviewers are positive: 73% give it 4 or 5 stars.
Many 5-star reviewers say PayPal payments arrive reliably and on time.
Reviewers value working on their own schedule, a few hours at a time, and many call the pay fair for the work.

What people complain about

Many applicants say they heard nothing for months after the one-shot Starter Assessment.
Recent reviewers say their accounts were blocked or locked without explanation, some with earnings still unpaid, and that support did not reply.
Work comes and goes: accepted workers describe empty dashboards and dry spells that last months.
The BBB rates DataAnnotation.tech F and lists 15 complaints in the last three years; it is not BBB accredited.

What has gone wrong

May 2025. A proposed class action in San Francisco Superior Court against 'Surge Labs, Inc. dba DataAnnotation' and Edwin Chen says annotators are misclassified as contractors, are not paid for onboarding assessments and qualification exams, and lose pay on tasks that run past time limits. Chen told Forbes the suit is without merit; it was ongoing in September 2025.
July 2025. Inc. reported that Surge AI had left a training document with safety guidelines for its contract workers open on public Google Docs. Surge unpublished it after Inc. asked and called it an old document written for internal research.

Who it suits

Best for: Fluent English writers, coders and credentialed experts who want self-scheduled remote work paid through PayPal, and who can treat it as extra income rather than a steady job. Not for: Anyone who needs guaranteed hours, a reply from support or a reason when an account is closed, or who cannot receive money through PayPal.

Honest catches

Pay here is platform-stated, not measured live from a job feed, so treat it as a benchmark rather than a guaranteed rate.

Sources for this review

What DataAnnotation pays

Platform-stated · as of 2026-07-25

$15–$100+/hr

Researched rate card, platform-stated, as of 2026-07-25

Rates from DataAnnotation's own FAQ and job board (read 15 Oct 2026): general $25–$50+/hr, coding $40–$150+/hr, STEM and professional $40–$125+/hr, languages up to $25–$40/hr, all with 'higher rates based on performance.' Independent 2026 worker reports put the effective general-task rate lower, often around $14–$20/hr. No public job feed.

Role	Kind of work	Researched pay
Coding / software engineering	Coding	$50–$100+/hr
STEM expert (math, physics, bio)	Math & STEM	$50–$100+/hr
Professional expert (law, finance, medicine)	Domain Experts	$50–$100+/hr
Generalist / AI Trainer	AI Evaluation	$25–$30+/hr

See how researched rate cards are compiled and graded.

Live DataAnnotation roles

DataAnnotation does not publish a public job feed.

You apply on DataAnnotation directly, then get matched to work from your account.

Join DataAnnotation

Some links on AITraining.jobs are referral links. If you sign up through one, AITraining.jobs may earn a commission at no extra cost to you. It never changes what a platform pays you.

Get new AI Evaluation roles by email

DataAnnotation has no public job feed we can watch, so we'll email you new AI Evaluation roles from the platforms that do, with real pay. Free, unsubscribe anytime.

Similar roles hiring now

DataAnnotation has no public job board. These live roles hire for the same kind of work on other verified platforms, refreshed hourly.

PhysioNet Competitor (Biomedical Signal Processing & ML) (opens the job details)
micro1 · AI Evaluation · Remote· 1d agoDetails
$60–$120/hr
ML Research Expert — Published First Author (Benchmark Authoring) (opens the job details)
AfterQuery · AI Evaluation · Remote· 1mo agoDetails
$150/hr
Staffing Supply Coordinator (opens the job details)
Turing · AI Evaluation · Remote· 6d agoDetails
Spanish Team Lead (opens the job details)
SuperAnnotate (SME) · AI Evaluation · Remote· 1mo agoDetails
up to $30/hr
AI Benchmark Quality Reviewer (opens the job details)
Braintrust · AI Evaluation · Remote· 4d agoDetails
$23/hr
Machine Learning Engineer – ML Evaluation & Experiment Design (opens the job details)
Anyone AI · AI Evaluation · Remote· 1mo agoDetails
$65/hr

Every site like DataAnnotation, compared

Some links on AITraining.jobs are referral links. If you sign up through one, AITraining.jobs may earn a commission at no extra cost to you. It never changes what a platform pays you.

The work DataAnnotation hires for

Each links to what the work involves, who hires for it, and what it pays across the market.

AI Evaluation & Rating (RLHF)Coding & Software EvaluationMath & STEMLaw, Medicine & FinanceWriting & ContentMultilingual & Localization

New to this? Start with the learning guides or compare pay across every platform.

How to get started on DataAnnotation

Researched August 2026

Sign up at the worker portal, immediately complete the one-shot Starter Assessment, and if your work is judged high-quality you are approved (usually within a few days to a few weeks), then pass short unpaid qualification tests that open paid projects.

Who can join

Be fluent in English. DataAnnotation states all workers must be fluent in English, and strong written English is the core skill for general tasks.
Have a bachelor's degree or equivalent real-world experience for general work. STEM projects want an advanced degree (master's/PhD) or a bachelor's plus 10+ years of professional experience; law/finance/medicine tracks want licensed credentials; coding tracks require real programming experience (Python, JavaScript, HTML, C++, C#, SQL).
Have reliable internet access and a computer, and turn off any VPN or location-masking tools during sign-up and the assessment.
Have a valid PayPal account (payment is PayPal-only). Provide the exact email address associated with that PayPal account.
Be located in an eligible country. DataAnnotation does not officially publish its country list; it is widely reported to be limited to a handful of English-speaking countries (commonly cited as the United States, Canada, United Kingdom, Ireland, Australia, and New Zealand), but third-party sources conflict on this, so confirm current eligibility on the sign-up page itself.

Step by step

1
Sign up at the worker portal
Create an account at the DataAnnotation worker sign-up page. Provide accurate details, complete your whole profile, and turn off any VPN so your location is not masked. Incomplete forms or VPN/location mismatches are reported to trigger instant rejection.
2
Take the Starter Assessment (one attempt only)
Right after signup you complete a take-at-home Starter Assessment: for the general track, rating and comparing AI chatbot responses and writing clear rationales, plus writing/grammar/reasoning tasks. You cannot retake it, so read instructions carefully and review before submitting.
3
Get approved
DataAnnotation reviews your submission for thoroughness and accuracy. Approval notification typically arrives within a few days, though worker guides report it can range from under 24 hours to several weeks. Rejection is usually silent (no email).
4
Complete the Core Assessment and unpaid Qualification Tests
According to worker guides (not the official FAQ), stronger candidates get a harder Core Assessment (ranking AI outputs with detailed reasoning, optional coding section), then short, unpaid domain-specific Qualification Tests. Passing more of these matches you to more, and better-paid, projects.
5
Start paid work and get paid via PayPal
Once qualified, projects appear in your dashboard based on your expertise and demonstrated quality. You log time on tasks and request payouts, which are delivered by PayPal within a few days.

Starter Assessment (worker guides describe a further Core Assessment and unpaid Qualification Tests)

A one-time, take-at-home evaluation you complete right after signing up. The general-track Starter Assessment tests writing, grammar, spelling, reading comprehension, and reasoning: typically rating or comparing AI chatbot responses and writing clear rationales for your judgments. Specialized tracks (coding, math, science, finance, law, medicine, language-specific) add domain tasks. DataAnnotation's own FAQ states you can only take the Starter Assessment once, with no retakes. Worker-facing guides (not the official FAQ) report that stronger candidates then face a harder Core Assessment (ranking AI responses with detailed reasoning, optional coding section) followed by short, unpaid domain Qualification Tests that unlock specific paid projects.

Format

Unproctored, take-at-home written assessment done in your browser. Specialized/coding assessments run about one to two hours per DataAnnotation; general assessments are shorter, though many applicants report spending 2+ hours. No live interview and no video call are mentioned in the process.

How to prepare

Read every instruction slowly before starting, and re-read your answers before submitting. It is one attempt with no retake, so DataAnnotation says to review carefully before submitting.
Optimize for quality and consistency, not speed. Reviewers look at thoroughness and accuracy, not how fast you finish.
On rating/comparison tasks, apply the rubric the same way every time and write specific, concrete reasons for your judgment. Vague reasoning is a commonly cited failure cause.
Write in clean, fluent English with correct grammar and spelling, and proofread. This is what general tasks pay for.
Do NOT use ChatGPT or other AI tools to generate your answers. Worker guides report this very likely results in an automatic fail, since the assessment exists to test human judgment against AI output.
Complete your profile fully and honestly, and pick the track (general vs coding) that matches your real strengths.

Why people get rejected

Inconsistency: applying the scoring rules differently across tasks.
Rushing and skimming instructions, so you miss formatting or task requirements.
Vague or thin rationales on evaluation tasks instead of specific, logical reasoning.
Using AI tools to write answers.
Spelling and grammar errors, or weak written English on a role whose core skill is writing.
Using a VPN / location-masking or mismatched location details, and incomplete application forms.

How long it takes

Approval after the Starter Assessment typically arrives within a few days per DataAnnotation, but worker guides report the realistic full range to get accepted is about 1 to 4 weeks, and occasionally longer (some applicants report waiting months with no reply). Rejection is usually silent. After you qualify and start paid projects, PayPal payouts are delivered within a few days after you request them.

When you get paid

Payment is PayPal only. Make sure the email on your DataAnnotation account exactly matches your PayPal email. DataAnnotation lists pay of about $25-$30+/hour for general work, $20+/hour for multilingual work, and $50-$100+/hour for coding, STEM, and professional projects. Deposits are delivered within a few days after you request a payout.

Tips to get accepted

Treat the Starter Assessment as a one-shot final exam: there are no retakes on your account, so do not start it until you can give it uninterrupted focus.
Be consistent, not just correct. Worker guides say applying the same rubric the same way every time matters more than raw correctness, so decide your logic and stick to it across all tasks.
Write specific rationales. On 'which response is better' tasks, explain exactly why (accuracy, instruction-following, tone), never a one-line 'it's better'.
Never paste AI-generated answers. Guides report the assessment screens for it and it very likely means an automatic fail, which defeats the point of a role that pays humans to evaluate AI.
Proofread ruthlessly for spelling and grammar. Clean written English is the actual product on general tasks.
Do not use a VPN or fudge your location, and complete your profile fully. These are cited as instant-rejection triggers.
After passing, keep taking qualification tests. More qualifications unlock more (and higher-paying) project queues, and steady high-quality work is what keeps tasks flowing.

What isn't public

DataAnnotation does not publish the exact rubric or scoring threshold for the assessments, so 'what counts as a pass' is inferred from worker reports, not disclosed.
The 'Core Assessment' and 'Qualification Tests' stage names do NOT appear in DataAnnotation's official FAQ; they come from worker-facing guides (e.g. OperationsArmy). The FAQ only names the Starter Assessment and its specialized variants.
Rejections are usually silent: the FAQ does not confirm a formal rejection email, and worker guides report you simply stop hearing back. There is no published way to see why you were rejected.
The FAQ does NOT list eligible countries or a minimum age. The commonly cited six-country list is not officially published and third-party sources conflict (some list additional regions), so it is a soft claim, not a quotable official one.
The general-track assessment content, exact wait times, and multi-stage structure are corroborated by worker guides rather than a single official flow chart, so timelines vary widely by person and demand for your skill set.
Whether a rejected applicant can ever reapply (new account) is not officially addressed; the only firm official rule is no retake on the same account.
The '$25-$30+/hour' and other pay figures are the platform's own stated ranges; actual earnings depend on task availability and how many qualifications you pass.

Start your DataAnnotation application

Sources (5)

Frequently asked questions

Is DataAnnotation legit?

Yes. DataAnnotation is the worker platform of Surge AI, a large, profitable San Francisco AI data company, and it pays approved work through PayPal. Your account or your work can vanish without a reason: reviewers describe locked accounts with earnings held, empty dashboards for months and support that rarely replies. The one-shot assessment and qualification tests are unpaid.

Is DataAnnotation a scam?

No. DataAnnotation is a real company. DataAnnotation warns that scammers use its name, says it never asks for money and that the only way to apply is its sign-up at app.dataannotation.tech. In a January 2025 BBB report, a fake 'Data Annotation Tech' task site on .cc lookalike domains, pitched by text and WhatsApp, took about $20,000 from a New Yorker. Never pay to get work.

Does DataAnnotation pay?

Yes. Pays approved work through PayPal and most Trustpilot reviewers are positive, but some describe accounts locked with earnings held and long gaps with no work; assessments are unpaid. Payouts: PayPal; paid within a few days of a payout request (its FAQ). Payment is PayPal only. Make sure the email on your DataAnnotation account exactly matches your PayPal email. DataAnnotation lists pay of about $25-$30+/hour for general work, $20+/hour for multilingual work, and $50-$100+/hour for coding, STEM, and professional projects. Deposits are delivered within a few days after you request a payout.

How much does DataAnnotation pay?

$15–$100+/hr. Rates from DataAnnotation's own FAQ and job board (read 15 Oct 2026): general $25–$50+/hr, coding $40–$150+/hr, STEM and professional $40–$125+/hr, languages up to $25–$40/hr, all with 'higher rates based on performance.' Independent 2026 worker reports put the effective general-task rate lower, often around $14–$20/hr. No public job feed.

Is DataAnnotation worth it?

DataAnnotation pays $15–$100+/hr (platform-stated). It rewards credentialed experts who clear its assessment far more than open generalist work. Worker sentiment: Trustpilot 3.9/5 (1,974 reviews).

How does DataAnnotation pay out?

PayPal; paid within a few days of a payout request (its FAQ).

Where can you work for DataAnnotation?

How do I actually get accepted onto DataAnnotation?

Sign up, then pass the one-time Starter Assessment by producing thorough, consistent, well-written answers. If your work is judged high quality you are approved (often within a few days), then you pass short unpaid qualification tests that open paid projects. No interview is mentioned in the process.

What is on the assessment and how hard is it?

The general Starter Assessment tests writing, grammar, reading comprehension, and reasoning, mostly by having you rate or compare AI chatbot responses and justify your judgment. It rewards following the rubric consistently and writing clearly rather than trick questions. Coding/STEM tracks add domain tasks and run about one to two hours.

Can I retake the assessment if I fail?

No. DataAnnotation's FAQ states you can only take the Starter Assessment once, with no retakes or second chances on your account. Read every instruction and review your answers before submitting.

Why do people get rejected?

Commonly cited reasons include inconsistency (applying the scoring rules differently across tasks), rushing and missing instructions, vague reasoning, using AI tools to write answers, spelling/grammar errors, and using a VPN or submitting an incomplete profile.

How long until I know, and until I get paid?

DataAnnotation says approval commonly comes within a few days; worker guides put the realistic range at 1-4 weeks (sometimes longer). Rejection is usually silent. Once you start paid work, PayPal payouts are delivered within a few days after you request them.