Quick answer: Yes, DataAnnotation.tech is legit — it’s the contributor-facing platform run by Surge AI, and it has paid contributors real money for years. The “is this a scam” searches mostly trace back to three real things: an unpaid starter assessment, task queues that dry up without warning, and a small but documented pattern of accounts getting deactivated with no explanation — sometimes with pay still pending. None of that makes it fake. It does mean you should treat it like the gig work it is, not a job offer.
Disclosure: This post may contain referral links. If you sign up through one, I may earn a bonus at no cost to you. It doesn’t change anything below — including the parts DataAnnotation wouldn’t put in a recruiter email.
In this guide
Who Actually Runs DataAnnotation.tech
DataAnnotation.tech is the contributor brand of Surge AI, a data-labeling company founded by Edwin Chen in 2020. Surge AI supplies training and evaluation data to AI labs building large language models — the actual client work behind the tasks you see in your dashboard. The company says it has paid out more than $20 million to contributors since launch, and it doesn’t charge anything to sign up or take the assessment. If a page or DM asks you to pay to “unlock” tasks, that’s not this platform — that’s someone impersonating it.
So Is It Legit? What the Evidence Actually Shows
Legitimacy and “will I have a good experience” are two different questions, and most of the online debate mixes them up. On the legitimacy question specifically:
- It pays. Weekly payouts via PayPal are the norm once you clear the minimum threshold, and there’s no pattern of mass non-payment across years of reviews.
- It has a real public track record. Roughly 1,900+ reviews on Trustpilot, mixed but leaning positive, plus an active Better Business Bureau profile — companies running scams don’t usually stick around long enough to accumulate either.
- The work is real. You’re comparing AI responses, writing model answers, or reviewing code — the same category of task Surge AI’s actual clients pay for.
Where it earns genuine skepticism is separate from all that — and worth being upfront about.
The Real Complaints (Not the Scam-Adjacent Ones)
Silent account deactivations. This is the most serious recurring issue. Contributors report accounts being deactivated for a vague “quality bar or Terms of Service” reason, with no specific explanation and no appeal path. One BBB complaint documented a contributor with over $5,500 in submitted work stuck in “pending approval” status after a deactivation, unable to get it released or even get a reason. This doesn’t happen to everyone, but it happens often enough that you shouldn’t let earnings sit in pending status any longer than necessary.
Task droughts. The dashboard can go empty for days or weeks with zero explanation, then refill. This is normal client-demand cycling in AI training work generally — it’s not unique to this platform — but it feels a lot like being ghosted if you don’t know that going in.
Thin, slow support. When something does go wrong, getting a human response can take a while, which amplifies every other frustration on this list.
No rejection feedback. If you don’t hear back within about two weeks of applying, you likely didn’t pass — but there’s rarely a formal rejection email telling you that outright.
The Pay, Honestly
Rates vary by task type, and self-reported figures across contributors are wide, but the general shape holds across sources:
- General tasks (writing prompts, ranking responses, basic evaluation): roughly $14–$20/hour
- Coding tasks (reviewing or writing code for AI training): roughly $25–$45/hour
- Domain-expert tasks (medicine, law, finance, other specialized fields): $25–$55+/hour, but these roles are competitive and availability fluctuates
Treat the top end of any range as the exception, not the baseline — the same rule that applies to every platform in this space.
How It Stacks Up Against the Alternatives
If you’re also weighing this against Outlier or want the wider international landscape, I’ve covered both in detail:
- Outlier AI Assessment 2026 — the closest direct comparison; different retake policy, different pay structure
- Best AI Training Platforms for Internationals — for readers outside the US/UK checking access first
Should You Apply?
If you want flexible, no-schedule side income and you’re comfortable with unpredictable task volume, this holds up as a legitimate option — the pay is real and the work is real. If you’re expecting anything close to steady full-time income, or you’re not prepared for the possibility of an unexplained deactivation, go in with that expectation set correctly rather than finding out the hard way.
Ready to actually apply? I’ve broken down exactly what the starter assessment tests and why most people fail it in my full starter assessment guide — worth reading before you touch the test, since you only get one attempt.
FAQ
Is DataAnnotation.tech a scam?
No. It’s a legitimate contributor platform run by Surge AI, with a multi-year track record of paying contributors and no cost to join. The frustrations people report — task droughts, slow support, occasional unexplained deactivations — are real, but they’re service-quality issues, not evidence of a scam.
How much does DataAnnotation.tech actually pay?
General tasks run roughly $14–$20/hour, coding tasks $25–$45/hour, and domain-expert work $25–$55+/hour, though rates and availability vary by project and fluctuate over time.
Why was my account deactivated with no explanation?
This is a documented pattern, not something you necessarily did wrong. Contributors report deactivations citing a general quality or Terms of Service reason with no specifics. If it happens to you, withdraw any available balance immediately and keep your own records of submitted work, since pending earnings can be difficult to recover.
Is the starter assessment paid?
No, it’s unpaid and you get one attempt on a given account — no retakes. See the full breakdown of what it tests in my starter assessment guide linked above.
