DataAnnotation (Surge Labs): Contributor Platform Profile, Pay and Reputation

DataAnnotation is a contributor platform at dataannotation.tech that pays people by the hour to write, rate and check responses used to train AI models, and the court record and its own site tie it to Surge Labs, Inc., the company behind Surge AI. DataAnnotation states that general projects start at USD 25 to 50 per hour and that expert rates reach up to USD 75 to 125+ per hour, checked 2026-11-09. It screens applicants with a one-time Starter Assessment, pays through PayPal on request, publishes a privacy notice but no terms of service, and faces a proposed misclassification class action, Cavalier v. Surge Labs, pending as of 2026-11-09.

Every pay figure on this page is DataAnnotation's own claim, checked on 2026-11-09, never a confirmed earning. DataPacts is an independent reference on AI training data licensing and the human work behind it: it takes no fee from any platform, has joined no referral program yet and covers DataAnnotation whether or not it pays referrals.

Each claim below carries its source and check date, and how DataPacts sources and dates every figure is set out in the methodology.

What DataAnnotation Is and What Contributors Do There

DataAnnotation recruits contractors to evaluate AI responses and to do expert work in coding, STEM, medicine, law, finance and 43 languages, paid per hour. The work areas named on dataannotation.tech, checked 2026-11-09, are these:

• General AI response evaluation: rating and comparing chatbot answers.
• Coding: writing and checking code that models produce.
• STEM: checking answers in science, technology, engineering and mathematics.
• Medical and legal work: expert review in medicine and law.
• Finance: expert review of financial reasoning and calculations.
• Language work: tasks in 43 languages.

The platform recruits and pays the contributors, and AI labs buy the human data that results. Each role card on the job board names a field and a stated maximum hourly rate, so the work is organized by expertise rather than by task type. DataAnnotation does not name its clients on its pages, so which labs receive the work is not stated in the sources checked. The term "data annotation" is also the generic name for labeling data, a meaning this page does not develop.

DataAnnotation is not a buyer of company data. Unlike data labeling companies that also license data from companies, it recruits individual contributors and sells their work onward. A company holding records has no program to apply to here, and DataAnnotation's pages describe none.

Who Runs DataAnnotation: The Surge Labs Link on the Record

DataAnnotation's own pages name no legal entity, so the link to Surge Labs, Inc. rests on two records: a 2025 class action complaint that sues "Surge Labs, Inc. dba DataAnnotation", and a logo file on dataannotation.tech named "da-by-surge-2026". The table grades each piece of evidence by type, as of 2026-11-09.

Each piece proves something different. The complaint caption and its recruiting statement are a plaintiff's allegations, not findings of a court. The logo file name is the operator's own asset naming, which points to Surge but is not a legal statement. No DataAnnotation page names a contracting entity, and no terms of service are public, so a contributor cannot read on the site which company owes the pay. A San Francisco fictitious business name statement, a Delaware entity record or terms naming the entity would settle the dba; none of these is in the sources checked, so Surge's own confirmation is not stated here. The company databases and press summaries that rank for "who owns" DataAnnotation are not used as sources. Edwin Chen appears in the record only as an individual defendant named in the complaint filed 2025-05-20.

The entity matters because it is the party a contributor contracts with, is paid by and can bring a claim against. On other platforms the terms name it: Outlier's terms name Scale AI, Inc. and Smart Ecosystem, Inc. (dba Outlier AI), checked 2026-11-09. For DataAnnotation, the complaint caption is the only document in the sources checked that names an entity at all.

Surge AI presents itself as a business serving AI labs. Its homepage lists Products, Off-the-Shelf, Training Lab and Expert Workforce, with no program for licensing company data, checked 2026-11-09. On the same date its workforce page carried one contractor listing at USD 200 to 400 per hour, Surge AI's own claim. Revenue and funding figures for Surge AI are not stated here. Surge AI's business with AI labs is covered on its own profile of Surge AI.

DataAnnotation Pay, Dated: Stated Rates by Role

DataAnnotation states that general projects start at USD 25 to 50 per hour and that experts earn up to USD 75 to 125+ per hour, checked 2026-11-09. The table lists every pay and payment term DataAnnotation states, each as its claim, as of 2026-11-09.

"Up to" marks a ceiling and "start at" an entry point; neither is an earning. DataAnnotation publishes no figure for what contributors actually receive, how many hours of work are available or how many people are active, and no contributor-confirmed earnings dataset exists to test the rates. The 2025 complaint alleges that recruitment promised hourly rates of USD 20 to 40, an allegation dated 2025-05-20, while the rates stated on 2026-11-09 begin at USD 25. Both stay on the record with their dates; the gap between them shows how stated pay can move, not what any contributor was paid.

A worked comparison shows how far the ceilings sit apart. On 2026-11-09 the language cards state up to USD 40 per hour and the Software Engineer card up to USD 150+, so the highest language ceiling is less than a third of the coding ceiling. The general-projects entry range of USD 25 to 50 overlaps the language ceiling, which means a language contributor's stated maximum sits inside the general entry band rather than above it.

The spread inside DataAnnotation is wide: language roles stop at USD 40 per hour and the Software Engineer card at USD 150+, both stated on 2026-11-09. Surge AI's single workforce listing at USD 200 to 400 per hour, checked the same day, is a separate claim on a separate site and is not a DataAnnotation rate. These figures can be set beside pay for AI training work on other platforms, where the same dated-claim rule applies.

The DataAnnotation Job Board: 133 Roles by Field

On 2026-11-09 DataAnnotation's job board listed 133 roles, led by 43 language roles, 19 medical roles and 14 finance roles. The table counts the role cards by field, parsed from dataannotation.tech as of 2026-11-09.

Languages form the largest group, about a third of the board, and carry the lowest stated ceiling: all 43 language cards state up to USD 40 per hour (checked 2026-11-09). Medicine and finance follow, and together with healthcare operations, law, accounting and economics they make up the expert fields behind the USD 75 to 125+ headline. Coding has only five cards, but its Software Engineer card carries the highest stated ceiling on the board.

Grouped by kind of expertise, the 133 roles fall into five blocks:

• Languages: 43 roles.
• Professional fields (finance 14, legal 8, accounting 7, economics 4): 33 roles.
• Technical fields (engineering 12, STEM 7, coding 5, cybersecurity 4, data and ML 2): 30 roles.
• Health (medical 19, healthcare operations 6): 25 roles.
• General and design: 2 roles.

The grouping is DataPacts' arithmetic on DataAnnotation's own field labels, and the five blocks sum to 133. Only one card is a general role, so most of the board asks for a language or a professional qualification rather than general evaluation skill.

The count is a snapshot. A job board changes as projects open and close, so the 133 roles describe 2026-11-09 only, and a card on the board is not a promise of available work in that field.

Getting In: DataAnnotation's Starter Assessment as Stated

DataAnnotation screens applicants with a Starter Assessment that takes about an hour and can be taken only once, checked 2026-11-09. Applicants who pass may then take specialized assessments, which DataAnnotation states take 1 to 2 hours, and results arrive within a few days. A failed or abandoned Starter Assessment therefore cannot be repeated under DataAnnotation's stated rule. The single attempt is the gate to every role card on the board, including the language and expert roles.

Several points are not stated. DataAnnotation's pages do not say whether assessment time is paid, and they publish no acceptance rate, so this page gives none. The Cavalier complaint alleges that DataAnnotation did not pay for training and qualification time, which remains an allegation. Test content, preparation and the way other platforms screen are covered under how platforms test and onboard contributors, not here.

How DataAnnotation Pays: PayPal Deposits on Request, No Minimum Hours

DataAnnotation pays through PayPal and states that deposits "will be delivered within a few days after you request them", with no minimum hours, checked 2026-11-09. The contributor decides when to request a deposit, and the platform does not require a set number of hours per week. Pay is stated per hour on every role card, not per task.

What DataAnnotation does not publish matters as much. Its public pages give no payment schedule beyond "on request", no payout currency, no PayPal fee rule and no statement of which tax form it issues to contractors. No referral program was found on dataannotation.tech in the sources checked, so DataAnnotation offers no stated bonus for bringing in other contributors. Other platforms differ: Outlier states weekly payment on Tuesdays through PayPal, AirTM or ACH, checked 2026-11-09, which gives contributors a fixed date that DataAnnotation's on-request model does not.

What DataAnnotation Publishes and What It Does Not

DataAnnotation publishes a privacy notice but no public terms of service: the notice, effective 2026-06-23, allows the collection of facial recognition data from contractors. The table sets the published items against the missing ones, as of 2026-11-09.

The privacy notice names Euverify Ltd (Ireland) as DataAnnotation's EU representative and Euverify Ltd (UK) as its UK representative. A named representative gives people in the EU and UK a contact for data protection requests, and the facial recognition clause tells contractors that biometric data may be collected. The notice does not say in which step of the work such data is collected. ID checks and biometric clauses across platforms are compared under contributor privacy on human data platforms.

No public terms of service were found on the site, and DataAnnotation states no country rule on its public pages; its FAQ asks only for reliable internet access and fluency in English, checked 2026-11-10. A contributor therefore sees the contract terms only after joining. Terms decide classification, ownership of the work and any arbitration clause, and none of them can be read in advance, so this page states no term DataAnnotation has not published.

Is DataAnnotation Legit? What the Record Shows

DataAnnotation is a real, operating platform: it states its pay, pays through PayPal, publishes a privacy notice and can be traced to Surge Labs, Inc. through a court record, but it names no legal entity and publishes no contributor terms. The documented side is concrete. The rates are stated on its own pages (checked 2026-11-09), the privacy notice of 2026-06-23 names EU and UK representatives, and the operator appears in a court caption and in its own logo file name.

Other points remain unresolved or contested. DataAnnotation's pages name no legal entity and no public terms, the Cavalier complaint pleads misclassification and unpaid qualification time as allegations pending as of 2026-11-09, and no confirmed earnings data exists. Review sites and forum threads rank for the question, but they hold unchecked, self-selected reports, so DataPacts states no rating or review score for DataAnnotation.

Imitation offers are a separate question from the platform itself. The pages checked for this profile sit on dataannotation.tech and app.dataannotation.tech, and no specific scam using the DataAnnotation name is in the sources checked. A message offering DataAnnotation work from another domain does not come from those pages. Four points can be checked against the record without relying on reviews:

• The domain: the pages cited here are on dataannotation.tech and app.dataannotation.tech.
• The stated rates and their date: the FAQ and the job board as checked on 2026-11-09.
• The privacy notice: effective 2026-06-23, naming Euverify Ltd as EU and UK representative.
• The court record: Cavalier v. Surge Labs, filed 2025-05-20, pending as of 2026-11-09.

Warning signs and checks that apply to any platform are covered under how to check a platform.

Cavalier v. Surge Labs: The Misclassification Complaint Against DataAnnotation

Cavalier v. Surge Labs, Inc. dba DataAnnotation, filed on 2025-05-20 in the Superior Court of California, County of San Francisco, is a proposed class action alleging that DataAnnotation misclassified contributors as independent contractors and did not pay for training and qualification time. The plaintiff, Dominique DonJuan Cavalier II, sues Surge Labs, Inc. dba DataAnnotation, Edwin Chen and Does 1-50. Every claim below is an allegation; no court has ruled on any of them. The table lists the record as of 2026-11-09.

The case is pending as of 2026-11-09, and no ruling or settlement was found. The court's online docket could not be accessed for this check, so the status rests on the complaint and on the absence of any later record in the sources checked. Nothing in the record shows the case as settled, decided or dismissed. No order certifying a class was found either, so the class remains proposed and no settlement or award exists in the sources checked.

The complaint relies on California's ABC test, Cal. Lab. Code § 2775 et seq., codified by AB 5, amended by AB 2257 and in force since 2020-01-01. It presumes a worker in California is an employee unless the hiring entity proves all three prongs, which is why such suits turn on how the work is controlled and whether it is part of the company's usual business. The test and the federal and EU rules beside it are explained under how AI training contributors are classified. No other incident, breach or regulator complaint about DataAnnotation was found in the sources checked as of 2026-11-09.

Benefits, Risks, Downsides and Alternatives of Working on DataAnnotation

Working on DataAnnotation offers flexible hours and stated expert rates of up to USD 75 to 125+ per hour (checked 2026-11-09), but the terms stay unpublished until a contributor joins and the misclassification complaint is unresolved. The four lists below set out each side.

Benefits (as DataAnnotation states them, checked 2026-11-09)

• No minimum hours.
• Stated expert ceilings of up to USD 75 to 125+ per hour, and up to USD 150+ on the Software Engineer card (checked 2026-11-09).
• Breadth: 133 roles, including 43 language roles.
• PayPal deposits within a few days of a request.

Risks

• A pending misclassification complaint that alleges unpaid qualification time.
• A privacy notice that allows collection of facial recognition data from contractors.
• No public terms of service to read before joining.
• One Starter Assessment attempt, with no stated retake.

Downsides

• Contractor status, with no benefits stated on DataAnnotation's pages.
• No published country rule.
• Stated ceilings are not earnings, and available work is not stated.

Alternatives

• Other contributor platforms with their own terms, among them Outlier and Alignerr (profiled in the section below).
• Employment in the same field instead of contract work.
• Not taking AI training work at all.

The alternatives state different terms on the same date. Alignerr states an average of USD 80 per hour and listings of USD 40 to 120 per hour, with Persona ID checks and Hubstaff time tracking, checked 2026-11-09. Outlier states onboarding of 30 to 90 minutes and weekly payment. None of these figures ranks one platform above another; each is the platform's own claim.

How This Profile Is Compiled: Own Pages, Court Records, Check Dates

This profile uses DataAnnotation's own pages, Surge AI's own pages and the court record, each with its check date; review sites, forums, company databases and press summaries are not used as sources. DataAnnotation's site and FAQ were checked on 2026-11-09, and the complaint was read in full. Where a source is silent, the page says "not stated" rather than filling the gap. Figures from the home page, the FAQ and the job board are kept as DataAnnotation wrote them and are never averaged or rounded.

DataAnnotation Among AI Training Platforms and Surge AI's Other Business

DataAnnotation is one of three contributor platforms DataPacts profiles under human data, next to Outlier, run with Scale AI, and Alignerr, powered by Labelbox. Outlier's terms name Scale AI, Inc. and Smart Ecosystem, Inc. (dba Outlier AI), and Alignerr contracts through ALIGNERR LLC, both checked 2026-11-09. All three, and eleven more, are set side by side in AI training jobs and platforms compared.

• Outlier AI, run with Scale AI.
• Alignerr, powered by Labelbox.
• The companies behind human data for AI and the work they buy.
• Response ratings as RLHF data.
• Cavalier among contributor lawsuits against human data platforms.
• The 43 language roles beside AI training work by language.
• Cavalier in the dated log of data buyer incidents.
• The companies behind Outlier and DataAnnotation: Scale AI vs Surge AI.

More About DataAnnotation

Frequently Asked Questions About DataAnnotation

Who owns DataAnnotation?

DataAnnotation's pages name no owner; a 2025 complaint sues Surge Labs, Inc. dba DataAnnotation, and the site's logo file is named "da-by-surge-2026". The complaint, filed 2025-05-20, also alleges that Surge AI recruits through DataAnnotation. No filing by Surge confirming the dba was found in the sources checked.

Is DataAnnotation the same company as Surge AI?

Surge AI is the brand of Surge Labs, Inc., and the record shows DataAnnotation as its contributor recruiting brand. That rests on the complaint's caption and allegation and on the logo file name. DataAnnotation names no entity, and no Surge filing confirming the link was found.

Can you retake the DataAnnotation Starter Assessment?

No: DataAnnotation states the Starter Assessment can be taken only once, checked 2026-11-09. It takes about an hour. Specialized assessments of 1 to 2 hours follow for those who pass, with results within a few days.

Does DataAnnotation pay for the assessment?

DataAnnotation's pages do not state whether assessment time is paid. The Cavalier complaint alleges that DataAnnotation did not pay for training and qualification time. That is an allegation in a case pending as of 2026-11-09, not a finding.

Does DataAnnotation accept contributors outside the US?

DataAnnotation states no country rule on its public pages; its FAQ asks for reliable internet access and fluency in English, checked 2026-11-10. Its job board listed 43 language roles among 133 on 2026-11-09, and its privacy notice names EU and UK representatives, but neither is a country rule.

How is DataAnnotation income taxed?

DataAnnotation engages contributors as contractors, and its pages do not state which tax form it issues. It publishes no payout currency or fee rule either.

Filing rules for this kind of contract income are covered under taxes for AI training contributors, and this page states no threshold or rate.